Rail transit vehicle air conditioning unit intelligent operation and maintenance method and system based on big data

Abstract

The application discloses a kind of based on big data's rail transit vehicle air conditioning unit intelligent operation and maintenance method and system, it is related to equipment operation maintenance technical field, the method includes: according to the entering number of multiple nodes in vehicle operation process in big data acquisition, carrying number prediction is carried out, and carrying number distribution is obtained;According to carrying number distribution, the load analysis of multiple air conditioning units is carried out, obtains multiple load parameters, and multiple first aging parameters are obtained by analysis;Maintenance parameters of the track where target vehicle is located are collected, vibration analysis is carried out in combination with carrying number distribution, multiple vibration parameters are obtained, multiple second aging parameters are obtained by classification;According to multiple first aging parameters and multiple second aging parameters, aging parameters of multiple air conditioning units are processed and obtained, operation and maintenance optimization is carried out, multiple operation and maintenance frequencies are obtained, and the operation and maintenance of multiple air conditioning units are carried out.The application solves the technical problems that air conditioning unit maintenance strategy and actual equipment wear do not match in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment operation and maintenance, and in particular to an intelligent operation and maintenance method and system for rail transit vehicle air conditioning units based on big data. BACKGROUND

[0002] Rail transit is an important means of transportation in daily life. Rail transit vehicles are often equipped with air conditioners to provide passengers with a comfortable ride experience. As the core equipment to ensure passenger comfort, the operation and maintenance efficiency of rail transit vehicle air conditioning units directly affects the operating cost and system reliability.

[0003] Traditional maintenance strategies are mostly based on fixed periods or single operating condition parameters (such as operating time) for maintenance planning, which is difficult to cope with the differential aging problems caused by dynamic passenger flow changes, car load fluctuations, and track vibration coupling. During peak hours, air conditioning units are in a high-load state due to the sudden increase in the number of passengers, which can accelerate the degradation of equipment performance.

[0004] In the prior art, the maintenance frequency does not match the actual equipment wear and tear, which can cause problems such as excessive maintenance or fault lag in rail transit vehicle air conditioning units, resulting in low operation and maintenance efficiency and mismatch between air conditioning unit maintenance strategies and actual equipment wear and tear. SUMMARY

[0005] The present application provides an intelligent operation and maintenance method and system for rail transit vehicle air conditioning units based on big data, which is used to solve the technical problems of low operation and maintenance efficiency and mismatch between air conditioning unit maintenance strategies and actual equipment wear and tear in the prior art.

[0006] In view of the above problems, the present application provides an intelligent operation and maintenance method and system for rail transit vehicle air conditioning units based on big data.

[0007] In a first aspect, the intelligent operation and maintenance method for rail transit vehicle air conditioning units based on big data comprises: collecting the number of passengers entering a plurality of entering nodes during the operation of a target vehicle according to big data, predicting the number of passengers carried by a plurality of vehicle units in the target vehicle, and obtaining a passenger distribution, wherein the target vehicle is a rail transit vehicle;

[0008] According to the passenger distribution, the load analysis of a plurality of air conditioning units in a plurality of vehicle units is performed to obtain a plurality of load parameters, and a plurality of first aging parameters are analyzed and obtained;

[0009] Collecting maintenance parameters of the track where the target vehicle is located, combining the passenger distribution, performing vibration analysis of the plurality of vehicle units, obtaining a plurality of vibration parameters, and classifying a plurality of second aging parameters are obtained;

[0010] According to the plurality of first aging parameters and the plurality of second aging parameters, aging parameters of the plurality of air conditioning units are processed to obtain operation and maintenance optimization, and a plurality of operation and maintenance frequencies are obtained to perform operation and maintenance on the plurality of air conditioning units.

[0011] In a second aspect, the application provides a big data-based intelligent operation and maintenance system for air conditioning units of rail transit vehicles, comprising:

[0012] A load analysis module is configured to perform load analysis on the plurality of air conditioning units of the plurality of vehicle units according to the load distribution, and obtain a plurality of load parameters and a plurality of first aging parameters.

[0013] A load analysis module is configured to perform load analysis on the plurality of air conditioning units of the plurality of vehicle units according to the load distribution, and obtain a plurality of load parameters and a plurality of first aging parameters.

[0014] A vibration analysis module is configured to collect maintenance parameters of a track where the target vehicle is located, combine the load distribution, perform vibration analysis on the plurality of vehicle units, obtain a plurality of vibration parameters, and classify a plurality of second aging parameters.

[0015] A vibration analysis module is configured to collect maintenance parameters of a track where the target vehicle is located, combine the load distribution, perform vibration analysis on the plurality of vehicle units, obtain a plurality of vibration parameters, and classify a plurality of second aging parameters.

[0016] The one or more technical solutions provided in the application have at least the following technical effects or advantages:

[0017] The application provides a big data-based intelligent operation and maintenance method and system for air conditioning units of rail transit vehicles, which significantly improves the accuracy of air conditioning unit aging state evaluation and the adaptability of operation and maintenance strategies by fusing dynamic passenger flow prediction, vibration feature analysis, and multi-source data collaborative modeling. Compared with traditional methods, the technical solutions provided by the application significantly reduce the maintenance lag or resource waste of air conditioning units caused by load fluctuation and vibration coupling effect.

[0018] The application achieves the technical effects of multi-dimensional dynamic optimization of operation and maintenance frequency, improvement of air conditioning unit reliability, and reduction of maintenance cost of air conditioning units during the life cycle. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 A flowchart illustrating an intelligent operation and maintenance method for air conditioning units in rail transit vehicles based on big data, provided in an embodiment of this application;

[0021] Figure 2 A schematic diagram of the structure of an intelligent operation and maintenance system for air conditioning units of rail transit vehicles based on big data, provided in an embodiment of this application;

[0022] The components represented by each number in the attached diagram are explained below:

[0023] The module includes: passenger capacity prediction module 11, load analysis module 12, vibration analysis module 13, and operation and maintenance optimization module 14. Detailed Implementation

[0024] This application provides a big data-based intelligent operation and maintenance method and system for air conditioning units in rail transit vehicles, which addresses the technical problem of mismatch between existing air conditioning unit maintenance strategies and actual equipment wear and tear.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0026] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0027] Example 1, as Figure 1 As shown, this application provides an intelligent operation and maintenance method for air conditioning units in rail transit vehicles based on big data, the method comprising:

[0028] S100: Based on big data, collect the number of people entering multiple entry nodes during the operation of the target vehicle, predict the number of people carried by multiple vehicle units within the target vehicle, and obtain the distribution of the number of people carried. The target vehicle is a rail transit vehicle.

[0029] In rail transit, the traditional passenger flow monitoring usually only counts the total passenger flow of the whole train or station, and cannot capture the dynamic distribution characteristics of passengers in different carriages. Due to the spatial difference of passenger boarding and alighting behavior, simply relying on total passenger flow for analysis will lead to deviation in the estimation of the actual carrying capacity of each carriage, thereby affecting the accuracy of subsequent equipment load analysis. The existing technology is difficult to realize fine prediction at the carriage level.

[0030] The step S100 in the method provided by the embodiment of the present application comprises:

[0031] According to big data, the total number of passengers entering multiple nodes during the running process of the target vehicle in multiple running time periods is collected, and multiple total entering numbers are obtained.

[0032] According to the multiple total entering numbers, the carrying capacity of multiple vehicle units in the target vehicle is predicted, and multiple carrying capacities are obtained as a carrying capacity distribution, wherein each carrying capacity comprises the average carrying capacity in the vehicle unit at each time in the running time period.

[0033] According to the multiple total entering numbers, the carrying capacity of multiple vehicle units in the target vehicle is predicted, and multiple carrying capacities are obtained, comprising:

[0034] According to rail transit operation data, multiple sample total entering number sets are collected, and the carrying capacity of multiple vehicle units under different sample carrying number distributions is collected, and a sample carrying capacity distribution set is labeled and obtained.

[0035] Using multiple total entering numbers as input features and using carrying capacity distribution as output features, a carrying capacity predictor is constructed based on machine learning.

[0036] Using training data, the carrying capacity predictor is iteratively trained to convergence using the multiple sample total entering number sets and the sample carrying capacity distribution set, wherein the training data comprises the multiple sample total entering number sets and the sample carrying capacity distribution set.

[0037] Illustratively, according to big data, the total number of passengers entering multiple nodes during the running process of the target vehicle in multiple running time periods is collected. Using the rail transit gate information big data at the time of entry, the total number of passengers entering the station during the running process of the rail transit vehicle is counted, and multiple total entering numbers of multiple stations and gates are obtained. Passengers entering from different gates will enter different vehicle units, i.e. carriages, according to certain rules, for example, passengers entering from the left gate will probably enter the left carriage in the direction of entry.

[0038] According to the big data of rail transit operation, a plurality of sample total entering person sets are collected, for example, 30 people enter the station through A gate and 20 people enter the station through B gate at 9 o'clock in the morning. The carrying person numbers in a plurality of vehicle units under different sample carrying number distributions are collected, the carrying person numbers in each vehicle unit are labeled, the labels are the car numbers and the carrying number distribution (for example, A1 car, 20 people; A2 car, 15 people; A3 car, 15 people), and a sample carrying number distribution set is obtained.

[0039] Exemplarily, a deep neural network is used to construct the carrying number predictor. The deep neural network (DNN) is a neural network that can imitate the structure and working principle of the neural network of the human brain, and can realize a high-performance solution to a complex task through hierarchical feature learning and weight adjustment. Exemplarily, the carrying number predictor is constructed, including an input layer, a hidden layer 1, a hidden layer 2 and an output layer. The input layer is used to receive the total entering person data, the hidden layer 1 has 256 nodes and uses a ReLU function for activation to extract high-order nonlinear features, the hidden layer 2 has 128 nodes and uses a ReLU function for activation to learn the passenger flow distribution pattern, the Dropout rate is set to 0.2 to prevent overfitting, and the output layer is used to output the carrying number distribution. The mean square error is used as the loss function. Exemplarily, a plurality of total entering person numbers are used as input features, the carrying number distribution is used as output features, the carrying number predictor is trained, the model weight is adjusted according to the output carrying number distribution, and the model converges until the accuracy rate of the output carrying number distribution reaches more than 95%, that is, the training of the carrying number predictor is completed.

[0040] Through the construction and training of the person number predictor, the total entering person number is converted into the carrying number which is difficult to count, the carrying person number in the car is finely predicted based on the entrance flow, and accurate and reliable data support is provided for load analysis.

[0041] S200: According to the carrying number distribution, the load analysis of a plurality of air conditioning units of a plurality of vehicle units is performed, a plurality of load parameters are obtained, and a plurality of first aging parameters are analyzed and obtained.

[0042] In rail transit, existing air conditioning unit load evaluation is mostly based on fixed rated power or simple threshold judgment, without considering the dynamic influence of actual passenger load fluctuation on air conditioning refrigeration / heat demand. Due to the large difference in carrying person numbers in different cars, the unified load evaluation method is easy to cause the overloading and aging of the unit in the high-load car to be underestimated, and the maintenance resources of the low-load car to be excessively consumed.

[0043] The step S200 in the method provided in the embodiment of the present application comprises:

[0044] According to the plurality of bearing people numbers in the bearing people number distribution, the mapping classification obtains a plurality of load parameters, wherein different bearing people numbers correspond to different load parameters;

[0045] The ratio of the plurality of load parameters to a preset load parameter of the air conditioning unit is calculated to obtain a plurality of load coefficients;

[0046] The plurality of load coefficients are used to correct a preset load aging parameter to obtain a plurality of first aging parameters.

[0047] In the embodiment of the application, according to the bearing people number distribution obtained according to the bearing people number prediction period output, the plurality of bearing people numbers are mapped and classified to each car to obtain a plurality of load parameters. Exemplarily, the bearing people number of each car and the load parameter (power, unit: kW) of the air conditioning unit are linearly mapped. Specifically, for example, when the load is 30 people, the load parameter of the air conditioning unit is 35 kW to ensure the temperature of the car, and when the load is 50 people, the load parameter of the air conditioning unit is 40 kW. The preset load parameter of the air conditioning unit is a load parameter of the air conditioning unit under normal operation at a moderate bearing people number, and the value range is between the minimum operating power and the maximum operating power. Exemplarily, the preset load parameter of the air conditioning unit is set to 25 kW.

[0048] Further, in the embodiment of the application, the ratio of the plurality of load parameters to the preset load parameter of the air conditioning unit is calculated to obtain a plurality of load coefficients. Load coefficient = load parameter ÷ preset load parameter. Exemplarily, when the load is 30 people, the load coefficient = 35 ÷ 25 = 1.4. The load coefficient is greater than 1, indicating that the load of the air conditioning unit is large, and the aging degree increases quickly.

[0049] The preset load aging parameter refers to a parameter for measuring the aging rate of the air conditioning unit, and the value range is between 0 and 1. Exemplarily, the preset load aging parameter is set to 0.5%, that is, the air conditioning unit will fail and be damaged without maintenance for 200 days. The preset load aging parameter is corrected by using the obtained plurality of load coefficients to obtain a plurality of first aging parameters. First aging parameter = preset load aging parameter × load coefficient. Exemplarily, when the load is 30 people, the first aging parameter = 0.5 × 1.4 = 0.7%.

[0050] In the embodiment of the application, the mapping of the load parameter is based on the bearing people number distribution, which can quantitatively analyze the change of the working strength of the air conditioning unit under different passenger loads. By introducing the load coefficient to correct the preset aging parameter, the high-load area caused by passenger aggregation can be accurately identified, the aging evaluation of different cars with different strategies can be realized, and the maintenance resource mismatch problem caused by the same aging evaluation of all cars in the traditional method can be avoided.

[0051] S300: Collect the maintenance parameters of the track where the target vehicle is located, combine the number of passengers carried, and perform vibration analysis on the plurality of vehicle units to obtain a plurality of vibration parameters, and classify to obtain a plurality of second aging parameters.

[0052] The conventional vibration analysis only focuses on the state of the track itself or whether the vehicle is empty, and ignores the influence of the number of passengers on the vibration characteristics of the vehicle body. In actual operation, uneven passenger distribution will cause the change of the vehicle center of gravity, and the number of passengers will cause the change of the vehicle mass, which will produce coupled vibration effects with track irregularities, joint gaps and other parameters, making it difficult to accurately evaluate the actual influence of vibration on the aging of the air conditioning unit.

[0053] The step S300 in the method provided by the embodiment of the application comprises:

[0054] Collecting the maintenance parameters of the track where the target vehicle is located;

[0055] According to the maintenance parameters, respectively combining a plurality of carrying passenger numbers in the carrying passenger number distribution, vibration prediction of a plurality of vehicle units is performed to obtain a plurality of vibration parameters;

[0056] According to the maintenance parameters, respectively combining a plurality of carrying passenger numbers in the carrying passenger number distribution, vibration prediction of a plurality of vehicle units is performed to obtain a plurality of vibration parameters, comprising:

[0057] According to the operation and maintenance data of the rail transit vehicle, a sample maintenance parameter set of the track, a sample carrying passenger number set of each vehicle unit are collected, and the maximum vibration amplitude of the air conditioning unit under different sample maintenance parameters and sample carrying passenger numbers is collected, and a sample vibration parameter set is labeled and obtained;

[0058] Taking the maintenance parameters and the carrying passenger numbers as input features and taking the vibration parameters as output features, a vehicle vibration predictor is constructed based on machine learning;

[0059] The sample maintenance parameter set, the sample carrying passenger number set and the sample vibration parameter set are used as training data to supervise the training of the vehicle vibration predictor until convergence;

[0060] The maintenance parameters are respectively combined with a plurality of carrying passenger numbers in the carrying passenger number distribution, and input into the vehicle vibration predictor to obtain a plurality of vibration parameters.

[0061] The ratio of the plurality of vibration parameters to the preset vibration parameters is calculated to obtain a plurality of vibration coefficients;

[0062] The plurality of vibration coefficients are used to correct the preset vibration aging parameters to obtain a plurality of second aging parameters.

[0063] In the embodiments of the present application, the maintenance parameters of the track where the target vehicle is located are collected. For example, the track maintenance parameter is the average value of the height difference between each track measured during track maintenance, measured using a vernier caliper, in millimeters. The maintenance parameter reflects the smoothness of the track. The lower the maintenance parameter, the smoother the track, and the smaller the vibration amplitude of the rail transit vehicle during operation.

[0064] According to the historical operation and maintenance records of the rail transit vehicle, different sample maintenance parameters of the track at different times are collected to obtain a sample maintenance parameter set; a set of carrying capacities of each vehicle unit is collected to obtain a sample carrying capacity set; the maximum vibration amplitude of the rail transit vehicle air conditioning unit under different sample maintenance parameters and sample carrying capacities is collected to obtain a sample vibration parameter set. The vibration amplitude is measured using an acceleration sensor, and the obtained data is in units of mm / s², reflecting the vibration amplitude of the air conditioning unit.

[0065] Exemplarily, a gradient boosting decision tree and a deep neural network hybrid model are used to construct a vehicle vibration predictor. Gradient boosting decision tree (GBDT, Gradient Boosting Decision Tree) is an iterative decision tree algorithm that constructs a set of weak learners (trees) and adds the results of multiple decision trees as the final prediction output. Using gradient boosting decision tree and deep neural network hybrid model can take into account feature interaction modeling and nonlinear mapping capability. Exemplarily, the vehicle vibration predictor is constructed, including a feature interaction layer, an input layer, a hidden layer 1, a hidden layer 2, and an output layer. The feature interaction layer is a GBDT module, including 100 trees, a maximum depth of 6, a learning rate of 0.1, inputting maintenance parameters and carrying capacities, and outputting a 20-dimensional feature vector. The input layer is used to receive the GBDT output features, with 20 nodes. The hidden layer 1 has 64 nodes and uses a ReLU function for activation. The hidden layer 2 has 32 nodes and uses a ReLU function for activation. A Dropout rate of 0.2 is set to prevent overfitting. The output layer is used to output the vibration parameter, with 1 node. The loss function uses mean squared error as the loss function. Exemplarily, the sample maintenance parameter set and the sample carrying capacity set are used as input features, and the vibration parameter is used as output features to train the carrying capacity predictor. The model weight is adjusted according to the output vibration parameter until the model converges, for example, the input maintenance parameter set and the carrying capacity set, and the output vibration parameter accuracy rate reaches more than 95%, which is the vehicle vibration predictor training completion.

[0066] The maintenance parameters are combined with multiple carrying capacities in the carrying capacity distribution, input into the trained vehicle vibration predictor, and the predicted output is obtained.

[0067] The ratio of the plurality of vibration parameters to preset vibration parameters is calculated respectively to obtain a plurality of vibration coefficients. The preset vibration parameter refers to a vibration parameter of the air conditioning unit under normal operation with moderate carrying capacity and track maintenance parameters, and is exemplarily set to 50 mm / s2. The vibration coefficient = vibration parameter ÷ preset vibration parameter, and is exemplarily 1.5 when the vibration parameter is 75 mm / s2 and the preset vibration parameter is 50 mm / s2. When the vibration coefficient is greater than 1, it indicates that the vibration amplitude of the air conditioning unit is large, and the aging degree increases rapidly.

[0068] The plurality of vibration coefficients are used to correct the preset vibration aging parameter to obtain a plurality of second aging parameters. The preset vibration aging parameter refers to a parameter for measuring the operation condition of the air conditioning unit, and the value range is between 0 and 1. The preset vibration aging parameter is exemplarily set to 0.5%. The plurality of vibration coefficients are used to correct the preset vibration aging parameter to obtain a plurality of second aging parameters. The second aging parameter = preset vibration aging parameter × vibration coefficient. The second aging parameter is exemplarily 0.75% when the vibration coefficient is 1.5.

[0069] In the embodiment of the application, the vibration prediction model is constructed by fusing the track maintenance parameters and the real-time carrying capacity distribution. The model can analyze the influence mechanism of the track maintenance parameters and the passenger capacity change on the vibration intensity of the air conditioning unit, identify the abnormal vibration mode caused by the combination of overload or track defects, extract the vibration feature parameters strongly related to the real aging process, compensate for the limitations of single factor analysis, quantify the aging condition of the air conditioning unit caused by the track maintenance parameters and the real-time carrying capacity, and is beneficial to subsequent data processing and determination of the operation and maintenance frequency of the air conditioning unit.

[0070] S400: According to the plurality of first aging parameters and the plurality of second aging parameters, the aging parameters of the plurality of air conditioning units are processed to obtain a plurality of operation and maintenance frequencies, and the operation and maintenance of the plurality of air conditioning units is performed.

[0071] The traditional operation and maintenance strategy often uses a fixed cycle or a single index threshold to trigger maintenance, which cannot adapt to the non-uniform aging characteristics of the air conditioning unit caused by the coupling of multiple factors such as load capacity and vibration. This leads to delayed maintenance of some high-aging-risk units, and premature maintenance of low-risk units, and it is difficult to balance the overall operation and maintenance cost and equipment reliability.

[0072] The step S400 in the method provided in the embodiment of the application includes:

[0073] According to the plurality of first aging parameters and the plurality of second aging parameters, the aging parameters of the plurality of air conditioning units are processed to obtain a plurality of operation and maintenance frequencies, and the operation and maintenance of the plurality of air conditioning units is performed.

[0074] a plurality of first maintenance frequencies for maintaining the plurality of air conditioning units are randomly generated respectively;

[0075] a first maintenance fitness is processed by obtaining the plurality of first maintenance frequencies;

[0076] wherein the first maintenance fitness is processed by obtaining the plurality of first maintenance frequencies, comprising:

[0077] a first efficiency fitness is calculated by calculating the ratio of the plurality of first maintenance frequencies and a plurality of aging parameters respectively;

[0078] a first cost fitness is calculated by calculating the ratio of a preset maintenance frequency and the plurality of first maintenance frequencies respectively;

[0079] an absolute difference value range of the first maintenance frequency of the air conditioning unit of each two adjacent vehicle units is calculated, and the reciprocal of the mean value of the plurality of absolute difference value ranges is calculated as a first maintenance combination fitness;

[0080] the first maintenance fitness is calculated by the first efficiency fitness, the first cost fitness and the first maintenance combination fitness.

[0081] the maintenance frequencies are randomly set and the maintenance fitness is calculated for optimization, and after the optimization converges, the plurality of maintenance frequencies with the maximum maintenance fitness are obtained for the maintenance of the plurality of air conditioning units.

[0082] aging parameters of the plurality of air conditioning units are processed and calculated according to the plurality of first aging parameters and the plurality of second aging parameters. Specifically, the first aging parameters and the second aging parameters are averaged to obtain the aging parameters. The corresponding first aging parameters and the plurality of second aging parameters refer to the first aging parameters and the second aging parameters of the same air conditioning unit under the same load number and the same track maintenance parameters. Aging parameter=(first aging parameter+second aging parameter)÷2. For example, the first aging parameter is 0.7, the second aging parameter is 0.75, and the aging parameter=(0.7+0.75)÷2=0.725.

[0083] a plurality of first maintenance frequencies for maintaining the plurality of air conditioning units are randomly generated respectively, for example, a plurality of first maintenance frequencies are randomly generated in Python using the random module, with the unit of day / time, for example, 2 times / month.

[0084] The average of the ratios of the plurality of first maintenance frequencies to the plurality of aging parameters is calculated to obtain a first efficiency fitness. The ratio of the first maintenance frequency to the plurality of aging parameters = the first maintenance frequency ÷ the aging parameter. For example, the first maintenance frequency is 2 times per month, and the aging parameter is 0.725, so the ratio of the first maintenance frequency to the plurality of aging parameters = 2 ÷ 0.725 = 2.75. There are multiple ratios of the first maintenance frequency to the plurality of aging parameters, which are 2.75, 2.5, and 2.25, so the first efficiency fitness = (2.75 + 2.5 + 2.25) ÷ 3 = 2.5. The greater the maintenance frequency, the better the maintenance quality.

[0085] The ratio of the preset maintenance frequency to the plurality of first maintenance frequencies is calculated to obtain a first cost fitness. The preset maintenance frequency refers to the maintenance frequency of the air conditioning unit under the normal operation of the moderate number of passengers and track maintenance parameters, and the unit is times per day. The first cost fitness = the preset maintenance frequency ÷ the first maintenance frequency. For example, the preset maintenance frequency is 4 times per month, and the first maintenance frequency is 2 times per month, so the first cost fitness = 4 ÷ 2 = 2. Too high maintenance frequency will result in waste of maintenance resources, and the smaller the first cost fitness, the more optimized the maintenance frequency.

[0086] The absolute difference amplitude of the first maintenance frequency of the air conditioning unit of each two adjacent vehicle units is calculated, and the reciprocal of the average of the plurality of absolute difference amplitudes is calculated as the first maintenance combination fitness. For example, there are two adjacent carriages A1 and A2, the first maintenance frequency of A1 is 2 times per month, and the first maintenance frequency of A2 is 4 times per month. The absolute difference amplitude of the first maintenance frequency = |A1 first maintenance frequency - A2 first maintenance frequency| = |2 - 4| = 2. The reciprocal of the average of the plurality of absolute difference amplitudes is calculated as the first maintenance combination fitness. For example, the absolute difference amplitudes of the first maintenance frequency of multiple adjacent carriages are 1, 2, 2, and 1, so the first maintenance combination fitness = 1 ÷ the average of the absolute difference amplitudes = 1 ÷ [(1 + 2 + 2 + 1) ÷ 4] = 1.5. The closer the maintenance frequency of adjacent vehicle units, the higher the maintenance efficiency, and the greater the maintenance combination fitness.

[0087] The maintenance frequency is randomly set and the maintenance fitness is calculated to optimize, and after the optimization converges, the plurality of maintenance frequencies with the maximum maintenance fitness is obtained to perform the maintenance of the plurality of air conditioning units.

[0088] The first maintenance fitness is calculated based on the first efficiency fitness, the first cost fitness, and the first maintenance combination fitness. For example, the first efficiency fitness, the first cost fitness, and the first maintenance combination fitness are weighted and summed, with the first efficiency fitness having a weight of 0.5, the first cost fitness having a weight of 0.3, and the first maintenance combination fitness having a weight of 0.2. For instance, if the first efficiency fitness is 2.5, the first cost fitness is 2, and the first maintenance combination fitness is 1.5, then the first maintenance fitness = 0.5 × 2.5 + 0.3 × 2 + 0.2 × 1.5 = 2.15.

[0089] The operation and maintenance frequency is randomly set and the maintenance fitness is calculated for optimization. After optimization convergence, the multiple operation and maintenance frequencies with the highest maintenance fitness are obtained, and the operation and maintenance of the multiple air conditioning units are carried out accordingly. Particle Swarm Optimization (PSO) is used for optimization. PSO is a classic swarm intelligence algorithm. This algorithm is simple in principle, requires few parameters, and is easy to implement. It finds the global optimal solution through information interaction between individuals. In this embodiment, PSO is used, with the number of particles set to 50, the inertia weight initially set to 0.8, gradually reduced to 0.4 after convergence, the acceleration constants C1 and C2 set to 1.5, and the maximum number of iterations set to 100. The randomly generated multiple operation and maintenance frequencies and the calculated maintenance fitness are input into the algorithm for calculation. After iteration, the optimal solution is finally obtained, and the multiple operation and maintenance frequencies with the highest maintenance fitness are obtained, such as 2 times / month for carriage A1 and 3 times / month for carriage A2. The operation and maintenance of the multiple air conditioning units are carried out according to the calculated multiple operation and maintenance frequencies with the highest fitness.

[0090] In this embodiment, a multi-dimensional equipment aging assessment system is constructed by integrating load aging and vibration aging parameters. Based on dynamic aging parameters, an optimization algorithm is used to generate differentiated maintenance frequencies, giving higher maintenance priority to air conditioning units in high-load carriages while avoiding ineffective maintenance of air conditioning units in low-load carriages. Ultimately, this achieves precise resource allocation for on-demand maintenance, reducing overall maintenance costs while improving equipment availability.

[0091] Example 2, as Figure 2 As shown, based on the same inventive concept as the intelligent operation and maintenance method for air conditioning units of rail transit vehicles based on big data provided in Embodiment 1, this embodiment of the invention also provides an intelligent operation and maintenance system for air conditioning units of rail transit vehicles based on big data, including:

[0092] The passenger capacity prediction module 11 is used to collect the number of people entering multiple entry nodes during the operation of the target vehicle based on big data, and to predict the passenger capacity of multiple vehicle units within the target vehicle to obtain the passenger capacity distribution, wherein the target vehicle is a rail transit vehicle.

[0093] The load analysis module 12 is configured to perform load analysis on the air conditioning units of the vehicle units according to the load distribution, obtain a plurality of load parameters, and analyze a plurality of first aging parameters;

[0094] The vibration analysis module 13 is configured to collect maintenance parameters of the track where the target vehicle is located, perform vibration analysis on the vehicle units in combination with the load distribution, obtain a plurality of vibration parameters, and classify a plurality of second aging parameters.

[0095] The operation and maintenance optimization module 14 is configured to process aging parameters of the air conditioning units according to the first aging parameters and the second aging parameters, perform operation and maintenance optimization, obtain a plurality of operation and maintenance frequencies, and perform operation and maintenance on the air conditioning units.

[0096] In one embodiment, the load prediction module 11 is further configured to:

[0097] According to big data, the total number of entries of a plurality of entry nodes during the running of the target vehicle in a plurality of running time periods is collected to obtain a plurality of total entry numbers.

[0098] According to the total entry numbers, the load prediction of the vehicle units in the target vehicle is performed to obtain a plurality of load numbers as a load distribution, wherein each load number includes the average load number in the vehicle unit at each time in the running time period.

[0099] According to the total entry numbers, the load prediction of the vehicle units in the target vehicle is performed to obtain a plurality of load numbers, including:

[0100] According to the rail transit operation data, a plurality of sample total entry number sets are collected, and the load numbers in the vehicle units under different sample load distributions are collected to obtain a sample load distribution set.

[0101] A plurality of total entry numbers are used as input features, a load distribution is used as output features, a load predictor is constructed based on machine learning.

[0102] The load predictor is iteratively trained to convergence using the sample total entry number set and the sample load distribution set as training data, wherein the training data includes the sample total entry number set and the sample load distribution set.

[0103] In one embodiment, the load analysis module 12 is further configured to:

[0104] According to the multiple load-bearing numbers in the load-bearing number distribution, the mapping classification obtains multiple load parameters, wherein different load-bearing numbers correspond to different load parameters;

[0105] The multiple load parameters are calculated with preset load parameters of the air conditioning unit to obtain multiple load coefficients;

[0106] The multiple load coefficients are used to perform load correction on the preset load aging parameters to obtain multiple first aging parameters.

[0107] In one embodiment, the vibration analysis module 13 is further configured to:

[0108] Collect maintenance parameters of a track where the target vehicle is located;

[0109] According to the maintenance parameters, vibration prediction of multiple vehicle units is performed in combination with multiple load-bearing numbers in the load-bearing number distribution to obtain multiple vibration parameters;

[0110] According to the maintenance parameters, vibration prediction of multiple vehicle units is performed in combination with multiple load-bearing numbers in the load-bearing number distribution to obtain multiple vibration parameters, including:

[0111] According to the operation and maintenance data of the rail transit vehicle, a sample maintenance parameter set of the track, a sample load-bearing number set of each vehicle unit, and a maximum vibration amplitude of the air conditioning unit under different sample maintenance parameters and sample load-bearing numbers are collected, and a sample vibration parameter set is labeled and obtained;

[0112] Taking the maintenance parameters and the load-bearing numbers as input features and taking the vibration parameters as output features, a vehicle vibration predictor is constructed based on machine learning;

[0113] The sample maintenance parameter set, the sample load-bearing number set, and the sample vibration parameter set are used as training data to perform supervised training on the vehicle vibration predictor until convergence;

[0114] The maintenance parameters are input into the vehicle vibration predictor in combination with multiple load-bearing numbers in the load-bearing number distribution to obtain multiple vibration parameters.

[0115] The multiple vibration parameters are calculated with preset vibration parameters to obtain multiple vibration coefficients;

[0116] The multiple vibration coefficients are used to perform load correction on preset vibration aging parameters to obtain multiple second aging parameters.

[0117] In one embodiment, the operation and maintenance optimization module 14 is further configured to:

[0118] According to the plurality of first aging parameters and the plurality of second aging parameters, processing is performed to obtain aging parameters of the plurality of air conditioning units;

[0119] A plurality of first maintenance frequencies for maintaining the plurality of air conditioning units are randomly generated respectively;

[0120] Processing is performed to obtain a first maintenance fitness of the plurality of first maintenance frequencies;

[0121] The processing performed to obtain the first maintenance fitness of the plurality of first maintenance frequencies includes:

[0122] The ratio of the plurality of first maintenance frequencies to the plurality of aging parameters is calculated respectively to obtain a first efficiency fitness;

[0123] The ratio of a preset maintenance frequency to the plurality of first maintenance frequencies is calculated respectively to obtain a first cost fitness;

[0124] The absolute difference amplitude of the first maintenance frequency of the air conditioning unit of each two adjacent vehicle units is calculated, and the reciprocal of the mean value of the plurality of absolute difference amplitudes is calculated as a first maintenance combination fitness;

[0125] According to the first efficiency fitness, the first cost fitness and the first maintenance combination fitness, a first maintenance fitness is calculated.

[0126] The maintenance frequency is randomly set and the maintenance fitness is calculated for optimization, and after the optimization converges, the plurality of maintenance frequencies with the maximum maintenance fitness are obtained for the operation and maintenance of the plurality of air conditioning units.

[0127] In summary, the embodiments of the present application have at least the following technical effects:

[0128] The present application provides a rail transit vehicle air conditioning unit intelligent operation and maintenance method and system based on big data, which significantly improves the accuracy of air conditioning unit aging state evaluation and the adaptability of operation and maintenance strategy by fusing dynamic passenger flow prediction, vibration feature analysis and multi-source data collaborative modeling. Compared with the traditional method, the technical scheme provided by the present application significantly reduces the air conditioning unit maintenance lag or resource waste caused by load fluctuation and vibration coupling effect.

[0129] The present application achieves the technical effects of multi-dimensional dynamic optimization of operation and maintenance frequency, improvement of air conditioning unit operation and maintenance efficiency and reduction of air conditioning unit maintenance cost in life cycle.

[0130] It should be noted that the above-mentioned embodiment sequences of the present application are merely for description only, but not for representing the advantages and disadvantages of the embodiments. And the above-mentioned embodiments of the present specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0131] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0132] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that are within the scope of the present application should be included. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A big data-based intelligent operation and maintenance method for rail transit vehicle air conditioning units, characterized in that, The method comprises: According to big data, the number of people entering multiple entering nodes in the running process of a target vehicle is collected, the carrying capacity of multiple vehicle units in the target vehicle is predicted, and a carrying capacity distribution is obtained, wherein the target vehicle is a rail transit vehicle; According to the carrying capacity distribution, the load analysis of multiple air conditioning units of multiple vehicle units is performed, multiple load parameters are obtained, and multiple first aging parameters are analyzed and obtained; Maintenance parameters of the track where the target vehicle is located are collected, and vibration analysis of the multiple vehicle units is performed in combination with the carrying capacity distribution, multiple vibration parameters are obtained, and multiple second aging parameters are classified and obtained, including: Collecting maintenance parameters of the track where the target vehicle is located; According to the maintenance parameters, vibration prediction of multiple vehicle units is performed in combination with multiple carrying capacities in the carrying capacity distribution, respectively, and multiple vibration parameters are obtained, including: According to the operation and maintenance data of the rail transit vehicle, a sample maintenance parameter set of the track, a sample carrying capacity set of each vehicle unit, and the maximum vibration amplitude of the air conditioning unit under different sample maintenance parameters and sample carrying capacities are collected, and a sample vibration parameter set is labeled and obtained; Taking the maintenance parameters and the carrying capacity as input features and taking the vibration parameters as output features, a vehicle vibration predictor is constructed based on machine learning; The sample maintenance parameter set, the sample carrying capacity set, and the sample vibration parameter set are used as training data to supervise the training of the vehicle vibration predictor until convergence; The maintenance parameters are input into the vehicle vibration predictor in combination with multiple carrying capacities in the carrying capacity distribution, respectively, and multiple vibration parameters are predicted and output; The ratios of the multiple vibration parameters to the preset vibration parameters are calculated, respectively, and multiple vibration coefficients are obtained; The multiple vibration coefficients are used to correct the preset vibration aging parameters, and multiple second aging parameters are obtained; According to the multiple first aging parameters and the multiple second aging parameters, the aging parameters of the multiple air conditioning units are processed, operation and maintenance optimization is performed, multiple operation and maintenance frequencies are obtained, and the operation and maintenance of the multiple air conditioning units are performed.

2. The big data-based intelligent operation and maintenance method for rail transit vehicle air conditioning units according to claim 1, characterized in that, According to big data, the number of people entering multiple entering nodes in the running process of a target vehicle is collected, the carrying capacity of multiple vehicle units in the target vehicle is predicted, and a carrying capacity distribution is obtained, including: According to big data, the total number of people entering multiple entering nodes in the running process of a target vehicle in multiple running time periods is collected, and multiple total entering numbers are obtained; According to the multiple total entering numbers, the carrying capacity of multiple vehicle units in the target vehicle is predicted, and multiple carrying capacities are obtained as a carrying capacity distribution, wherein each carrying capacity includes the average carrying capacity in the vehicle unit at each time in the running time period.

3. The big data-based intelligent operation and maintenance method for rail transit vehicle air conditioning units according to claim 2, characterized in that, According to the multiple total entering numbers, the carrying capacity of multiple vehicle units in the target vehicle is predicted, and multiple carrying capacities are obtained, including: According to rail transit operation data, multiple sample total entering number sets are collected, and the carrying capacity in multiple vehicle units under different sample carrying number distributions is collected, and a sample carrying capacity distribution set is labeled and obtained; The total number of entries is used as an input feature, the carrying capacity distribution is used as an output feature, and a carrying capacity predictor is constructed based on machine learning. The carrying capacity predictor is iteratively trained to convergence using the training data, which includes the multiple sets of sample total entry numbers and sample carrying capacity distribution sets.

4. The big data-based intelligent operation and maintenance method for rail transit vehicle air conditioning units according to claim 1, characterized in that, According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including 5. The big data-based intelligent operation and maintenance method for rail transit vehicle air conditioning units according to claim 1, characterized in that, According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including 6. The big data-based rail transit vehicle air conditioning unit intelligent operation and maintenance method according to claim 5, characterized in that, According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including According to the carrying capacity distribution, the load analysis of the air conditioning units of the multiple vehicle units is performed to obtain multiple load parameters, and multiple first aging parameters are analyzed and obtained, including The system is used to perform the method of any one of claims 1-6, and the system comprises: A carrying capacity prediction module is used to collect the number of entries at multiple entry nodes during the operation of a target vehicle based on big data, predict the carrying capacity of multiple vehicle units in the target vehicle, and obtain a carrying capacity distribution, wherein the target vehicle is a rail transit vehicle. A load analysis module is used to perform load analysis of multiple air conditioning units of multiple vehicle units based on the carrying capacity distribution, obtain multiple load parameters, and analyze and obtain multiple first aging parameters.

7. A big data-based intelligent operation and maintenance system for rail transit vehicle air conditioning units, characterized in that, A load analysis module is used to perform load analysis of multiple air conditioning units of multiple vehicle units based on the carrying capacity distribution, obtain multiple load parameters, and analyze and obtain multiple first aging parameters. ​ ​ The vibration analysis module is configured to collect maintenance parameters of a track where the target vehicle is located, combine the load distribution, perform vibration analysis on the vehicle units, obtain vibration parameters, and classify to obtain second aging parameters, including: Collecting maintenance parameters of a track where the target vehicle is located; According to the maintenance parameters, combining the load distribution, performing vibration prediction on the vehicle units, obtaining vibration parameters, including: According to the operation and maintenance data of the rail transit vehicle, a sample maintenance parameter set of the track, a sample load set of each vehicle unit, and the maximum vibration amplitude of the air conditioning unit under different sample maintenance parameters and sample load are collected, and a sample vibration parameter set is labeled and obtained; Taking the maintenance parameters and the load as input features and taking the vibration parameters as output features, a vehicle vibration predictor is constructed based on machine learning; The sample maintenance parameter set, the sample load set, and the sample vibration parameter set are used as training data to supervise the training of the vehicle vibration predictor until convergence; The maintenance parameters are combined with the load distribution to input the vehicle vibration predictor, and vibration parameters are obtained by prediction output; The vibration parameters are calculated to obtain vibration coefficients; The vibration coefficients are used to correct the preset vibration aging parameters to obtain second aging parameters; The operation and maintenance optimization module is configured to process the aging parameters of the air conditioning units according to the first aging parameters and the second aging parameters, perform operation and maintenance optimization, obtain operation and maintenance frequencies, and perform operation and maintenance on the air conditioning units.

Images