Air conditioner energy consumption data supervision method and system based on big data

Abstract

The invention relates to the technical field of air conditioner maintenance, and discloses an air conditioner energy consumption data supervision method and system based on big data, and the method comprises the steps: firstly collecting historical operation data of a central air conditioner control system, including energy consumption, operation setting, production parameters, environmental parameters and people flow records; and feature construction is performed on the data to form a feature data set, and a machine learning algorithm is utilized to construct an energy consumption prediction model of each central air conditioner. The model comprehensively considers multi-dimensional factors and predicts the energy consumption trend of the central air conditioner in a period of time in the future. By collecting the energy consumption data of the central air conditioner in real time and comparing the energy consumption data with the predicted data, the energy consumption abnormal point can be found in time, and the early warning signal is generated and sent to the air conditioner management personnel, so that the energy consumption of the central air conditioner system is effectively supervised, the operation cost is reduced, and the energy consumption management effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning maintenance, and in particular to a method and system for monitoring air conditioning energy consumption data based on big data. Background Art

[0002] In large public buildings such as shopping malls, central air-conditioning systems are important energy-consuming equipment, and their energy consumption supervision is of great significance for achieving energy conservation and emission reduction and reducing operating costs.

[0003] In the existing field of air conditioning energy consumption data supervision, it usually relies on traditional energy consumption monitoring systems and manual inspection methods. Traditional energy consumption monitoring systems can only provide basic energy consumption data, and lack comprehensive consideration of multi-dimensional factors such as environmental parameters, changes in human flow, and equipment status. As a result, it is impossible to accurately predict the future energy consumption trend of central air conditioners, and it is impossible to detect energy consumption anomalies in a timely manner. The manual inspection method is inefficient and cannot achieve real-time supervision of a large number of central air conditioning systems. At the same time, due to the subjectivity and discontinuity of manual inspections, it is difficult to detect short-term and sudden energy consumption anomalies, resulting in poor energy consumption management effects. To this end, we propose an air conditioning energy consumption data supervision method and system based on big data. Summary of the invention

[0004] The purpose of the present invention is to provide an air conditioning energy consumption data monitoring method and system based on big data to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: an air conditioning energy consumption data monitoring method based on big data, the method comprising:

[0006] S1: Collect historical operation data of the central air-conditioning control system, including energy consumption data, operation settings, production parameters of each central air-conditioning unit, as well as environmental parameters and human flow records during operation;

[0007] S2: construct features for the data collected in step S1 to form a feature data set;

[0008] S3: Using machine learning algorithms, an energy consumption prediction model is constructed for each central air conditioner. The energy consumption prediction model is based on a feature data set and comprehensively considers environmental parameters, changes in human traffic, equipment status, and time factors to predict the energy consumption trend of the central air conditioner in the future.

[0009] S4: Obtain the current production parameters and operation settings of each central air conditioner, as well as the current environmental parameters and human flow data;

[0010] S5: input the acquired data into the energy consumption prediction model, and receive the energy consumption prediction trend data output by the model;

[0011] S6: Collect the real-time energy consumption data of each central air conditioner from the central air conditioning control system, and clean the collected raw data to remove abnormal values;

[0012] S7: Plot the real-time energy consumption data collected in step S6 and the predicted energy consumption data output by the model in step S5 into curves that change over time, align the timestamps of the two curve data, and determine whether each point on the curve is abnormal according to a preset algorithm, and if so, mark it as an abnormal point;

[0013] S8: Generate a warning signal based on the discovered abnormal points and warning rules, and further send the warning signal to the air conditioning management personnel.

[0014] Preferably, the feature data set includes:

[0015] Time window characteristics, including average energy consumption, maximum energy consumption, and minimum energy consumption in the past preset time period;

[0016] Crowd flow-related features, including the ratio of current crowd flow to the average crowd flow in a preset time period in the past, the rate of change of crowd flow, and the identification of crowd flow during peak hours;

[0017] Environmental parameter characteristics, including indoor and outdoor temperature difference, humidity change rate, and whether it is in extreme weather conditions;

[0018] Equipment status characteristics, including equipment operation time, time since the last equipment maintenance, and whether the equipment is under maintenance;

[0019] Time-related features, including the day of the week, day or month of the year, and identification of holidays;

[0020] Statistical characteristics, including the distribution characteristics of energy consumption data and the rolling statistics of energy consumption data;

[0021] Rate of change characteristics, including the rate of change of energy consumption.

[0022] Preferably, a long short-term memory network (LSTM) algorithm is used to construct an energy consumption prediction model, and the specific steps include:

[0023] S301: Using a feature data set as input, dividing the feature data set into a training set and a test set;

[0024] S302: Build the LSTM network structure, set the number of network layers, the number of hidden layer units and the activation function hyperparameters;

[0025] S303: Use the training set to train the LSTM model, and optimize the model parameters through the back propagation algorithm and the gradient descent method;

[0026] S304: Use the test set to evaluate the trained LSTM model and verify the prediction performance of the model;

[0027] S305: adjusting the hyperparameters of the LSTM model according to the evaluation results until a satisfactory prediction accuracy is achieved;

[0028] S306: Train and save the corresponding LSTM energy consumption prediction model for each central air conditioner.

[0029] Preferably, the network structure of the LSTM model is:

[0030] Input layer: receives feature data sets as input, which include time window features, crowd flow-related features, environmental parameter features, equipment status features, time-related features, statistical features, and change rate features;

[0031] Network layer: contains two hidden layers, each containing 128 and 64 hidden units respectively. Each hidden unit updates and transmits information through the forget gate, input gate and output gate, and uses the ReLU activation function to perform nonlinear transformation on the output of the hidden unit.

[0032] Output layer: The output layer receives the output of the last hidden layer and maps the output of the LSTM network to the energy consumption prediction value through a fully connected layer.

[0033] Preferably, the LSTM model is iteratively trained using the training set data. In each iteration, a batch of input data is sent to the LSTM model, and the memory units and hidden states are updated by calculating the values ​​of the input layer and the output layer, and finally the predicted output sequence is obtained. During the training process, the validation set is regularly used to evaluate the model performance, and the learning rate is adjusted or the training is stopped early as needed to prevent overfitting.

[0034] Preferably, the mean square error MSE algorithm is used to evaluate the performance of the energy consumption prediction model. The specific algorithm is:

[0035]

[0036] Where n is the number of data points, Y pred,i is the energy consumption value predicted by the model, Y actual,i is the actual measured energy consumption value; the prediction accuracy of the model is evaluated according to the MSE value. The smaller the MSE value, the higher the prediction accuracy of the model.

[0037] Preferably, the method of cleaning the original data in step S6 includes: setting a reasonable value range for each type of data, traversing the original data, marking the values ​​exceeding the set range as outliers, and removing the data marked as outliers from the original data.

[0038] Preferably, the algorithm for marking outliers includes:

[0039] Difference threshold-based algorithm: Set a fixed energy consumption difference threshold. When the difference between the real-time energy consumption data and the predicted energy consumption data exceeds the threshold, it is marked as an abnormal point. The calculation formula is: if |RP|>T, it is marked as an abnormal point, where R represents the real-time energy consumption data, P represents the predicted energy consumption data, and T represents the set energy consumption difference threshold;

[0040] Algorithm based on percentage of difference: set a percentage threshold of energy consumption difference to predicted value. When the percentage of difference between real-time energy consumption data and predicted energy consumption data to predicted value exceeds the threshold, it is marked as an abnormal point. The calculation formula is: if |(RP) / P|>PT, it is marked as an abnormal point, where R represents real-time energy consumption data, P represents predicted energy consumption data, and PT represents the set percentage threshold of energy consumption difference to predicted value.

[0041] Algorithm based on sliding average: Calculate the average value of energy consumption difference within a time window. When the difference between real-time energy consumption data and predicted energy consumption data exceeds a certain multiple of the average value, it is marked as an abnormal point. The calculation formula is: if |RP|>M×AVG, it is marked as an abnormal point, where R represents real-time energy consumption data, P represents predicted energy consumption data, M represents the set multiple, and AVG represents the average value of energy consumption difference within the time window;

[0042] Algorithm based on trend change: Analyze the changing trends of real-time energy consumption data and predicted energy consumption data. When the trend of real-time data deviates significantly from the trend of predicted data, it is marked as an abnormal point. The trend change is realized by calculating the slopes of the two curves. The formula for the slope difference is: if |Slope(R)-Slope(P)|>ST, it is marked as an abnormal point, where Slope(R) represents the slope of real-time energy consumption data, Slope(P) represents the slope of predicted energy consumption data, and ST represents the set slope difference threshold.

[0043] Preferably, the warning signal includes a specific timestamp of the abnormal point, a corresponding real-time energy consumption value, a predicted energy consumption value, an abnormal type identifier, and the number and location information of the relevant central air conditioner.

[0044] Preferably, an air conditioning energy consumption data monitoring system based on big data comprises:

[0045] The data collection module is responsible for collecting the historical operation data of the central air-conditioning control system, including the energy consumption data, operation settings, production parameters of each central air-conditioning unit, as well as the environmental parameters and human flow records during operation;

[0046] The feature construction module constructs features of the collected data to form a feature data set for subsequent modeling;

[0047] The model building module uses machine learning algorithms to build an energy consumption prediction model for each central air conditioner. Based on the feature data set, the model comprehensively considers environmental parameters, changes in human traffic, equipment status, and time factors to predict the energy consumption trend of the central air conditioner in the future.

[0048] The real-time data acquisition module is responsible for obtaining the current production parameters and operation settings of each central air conditioner, as well as the current environmental parameters and human flow data;

[0049] The energy consumption prediction module inputs real-time data into the energy consumption prediction model and receives the energy consumption prediction trend data output by the model;

[0050] The data cleaning module collects the real-time energy consumption data of each central air conditioner from the central air conditioning control system and cleans the collected raw data to remove abnormal values;

[0051] The anomaly detection module plots the cleaned real-time energy consumption data and predicted energy consumption data into time-varying curves, aligns the timestamps of the two curve data, and determines whether each point on the curve has an anomaly based on a preset algorithm. If so, it is marked as an abnormal point.

[0052] The warning signal generation and sending module generates a warning signal based on the discovered abnormal points and preset warning rules, and sends the warning signal to the designated air conditioning management personnel.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] By acquiring the production parameters, operation settings, environmental parameters and traffic data of central air conditioners in real time and inputting them into the energy consumption prediction model, energy consumption anomalies can be discovered in a timely manner. Compared with manual inspection methods, this method can detect short-term and sudden energy consumption anomalies more quickly and accurately, so that timely measures can be taken to deal with them. This method realizes real-time supervision of a large number of central air conditioning systems and significantly improves the efficiency of energy consumption management. At the same time, by automatically generating early warning signals and sending them to air conditioning management personnel, the energy consumption management process is further simplified and the manual burden is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a step diagram of the air conditioning energy consumption data monitoring method based on big data according to the present invention;

[0056] Figure 2 Flowchart for building and training energy consumption prediction model;

[0057] Figure 3 This is a structural diagram of the air conditioning energy consumption data monitoring system based on big data described in the present invention. DETAILED DESCRIPTION

[0058] 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.

[0059] See also Figure 1-3 The present invention provides a technical solution: an air conditioning energy consumption data monitoring method based on big data, the steps of the method include:

[0060] Data collection stage (S1): First, historical operation data is comprehensively collected from the central air-conditioning control system. The historical operation data includes but is not limited to the energy consumption data, operation setting parameters, production parameters of each central air-conditioning unit, as well as environmental parameters (such as temperature, humidity) and human flow records during operation.

[0061] Feature construction stage (S2): The collected raw data is used to construct a feature data set through feature engineering technology, and features that are valuable for energy consumption prediction are extracted from the raw data.

[0062] Model building stage (S3): Using machine learning algorithms, an energy consumption prediction model is built for each central air conditioner. The energy consumption prediction model is based on a feature data set, taking into account environmental parameters, changes in human traffic, equipment status, and time factors (such as season, working day / non-working day). Through training and learning the historical energy consumption pattern of the central air conditioner, it is possible to predict the energy consumption trend in the future.

[0063] Real-time data acquisition stage (S4): The current production parameters and operation settings of each central air conditioner are acquired in real time, and the current environmental parameters and human flow data are recorded as part of the model input to ensure the accuracy of the prediction.

[0064] Energy consumption prediction stage (S5): the data obtained in step S4 is input into the corresponding energy consumption prediction model, and the model outputs energy consumption prediction trend data for a period of time in the future to provide a benchmark for subsequent anomaly detection.

[0065] Real-time energy consumption data collection and processing stage (S6): The energy consumption data of each central air conditioner is collected in real time from the central air conditioning control system, and the raw data is cleaned to remove abnormal values ​​caused by sensor failure and other reasons to ensure the reliability of the data.

[0066] Abnormal detection stage (S7): The real-time energy consumption data and the energy consumption data predicted by the model are plotted as time-varying curves, and the two curves are time-stamped and aligned. A preset algorithm (such as threshold-based detection, time series analysis, etc.) is used to determine whether each point on the curve has an abnormality. If an abnormality is detected, it is marked as an abnormal point.

[0067] Early warning signal generation and sending stage (S8): Early warning signals are generated based on the detected abnormal points and the preset early warning rules. These signals contain information such as the type, location, and severity of the abnormality. Subsequently, these early warning signals are sent to air conditioning managers in a timely manner so that they can quickly take measures to deal with potential energy consumption problems or equipment failures, thereby achieving effective energy consumption management and cost control.

[0068] In order to effectively monitor the energy consumption of the central air conditioning system and promptly detect and handle abnormal points, the early warning signal contains specific information about the abnormal points so that relevant personnel can quickly locate the problem and take corresponding measures. The following is a detailed description of the early warning signal content and specific implementation methods:

[0069] The early warning signal mainly includes the following parts: the specific timestamp of the abnormal point, the corresponding real-time energy consumption value, the predicted energy consumption value, the abnormal type identification, and the number and location information of the relevant central air conditioner.

[0070] Specific timestamp of the anomaly: This refers to the exact time when the anomaly occurred, usually expressed in the form of year, month, day, hour, minute, and second. It helps relevant personnel understand the time context of the anomaly and better analyze the cause of the anomaly.

[0071] Corresponding real-time energy consumption value: This refers to the actual energy consumption value of the central air conditioning system when the abnormal point occurs. By comparing the real-time energy consumption value with the predicted energy consumption value, the nature and possible causes of the abnormal point can be further analyzed.

[0072] Predicted energy consumption value: This refers to the predicted energy consumption value corresponding to the time when the abnormal point occurs based on the energy consumption prediction model. It provides a reference benchmark for comparison with the real-time energy consumption value to determine whether the energy consumption is abnormal.

[0073] Abnormal type identification: This is an identification for classifying abnormal points, such as "difference threshold abnormality", "difference percentage abnormality", "sliding average abnormality" or "trend change abnormality". It helps relevant personnel quickly understand the type and possible cause of the abnormal point, so as to take more targeted treatment measures.

[0074] The number and location information of the relevant central air conditioner: This refers to the number and location information of the central air conditioning system where the abnormality occurred, such as floor, room number, etc. It helps relevant personnel quickly locate the specific equipment and location where the problem occurs, so as to conduct further inspection and processing.

[0075] In specific implementation, when the energy consumption monitoring system detects an abnormal point, it will automatically generate an early warning signal and include the above information. For example, suppose that at 2:30 pm on May 1, 2023, the central air-conditioning system numbered 001 is located in the lobby on the first floor. Its real-time energy consumption value is 150KW·h, while the predicted energy consumption value is 100KW·h. At this time, the energy consumption monitoring system determines that the point is a "difference threshold abnormality" and automatically generates an early warning signal. The signal contains the following information: "Abnormal time: 2023-05-01 14:30; real-time energy consumption value: 150KW·h; predicted energy consumption value: 100KW·h; abnormal type: difference threshold abnormality; equipment number: 001; equipment location: lobby on the first floor". After receiving the early warning signal, relevant personnel can quickly locate the problem and take appropriate measures to deal with it.

[0076] The present invention will be further described below in conjunction with Examples 1 to 3:

[0077] Embodiment 1:

[0078] The feature data set covers information in multiple dimensions to ensure the comprehensiveness and accuracy of the energy consumption prediction model. Specifically, it includes the following features:

[0079] ①Time window characteristics:

[0080] Average energy consumption during a preset time period in the past: reflects the stable level of energy consumption over a period of time.

[0081] Maximum and minimum energy consumption during a preset time period in the past: reveals the fluctuation range of energy consumption.

[0082] Assuming that the preset time period is the past week, the average energy consumption in the past week can be calculated by adding the energy consumption data of each day in the week and dividing by 7. Similarly, the maximum energy consumption and the minimum energy consumption in the past week can be determined by finding the maximum value and the minimum value of the energy consumption data in the week.

[0083] ②Characteristics related to human traffic:

[0084] The ratio of current passenger flow to the average passenger flow in the past preset time period: reflects the relative change in passenger flow.

[0085] The rate of change of passenger flow: reflects the dynamic change trend of passenger flow.

[0086] Peak hour traffic flow identification: used to mark peak traffic hours so that the model can consider its impact on energy consumption.

[0087] The ratio of the current flow of people to the average flow of people in the past week can be calculated by dividing the current flow of people by the average flow of people in the same time period every day in the past week. The rate of change of the flow of people can be obtained by calculating the difference between the flow of people in the current time period and the previous time period, and then dividing it by the flow of people in the previous time period. The identification of the flow of people during peak hours can be done by setting a threshold. When the flow of people exceeds this threshold, it is marked as peak hours.

[0088] ③Environmental parameter characteristics:

[0089] Indoor and outdoor temperature difference: reflects the difference between indoor and outdoor environment and has a significant impact on energy consumption.

[0090] Humidity change rate: reflects the change of humidity over time and has a certain impact on energy consumption.

[0091] Flag for extreme weather conditions: used to mark extreme weather conditions so that the model can consider their special impact on energy consumption.

[0092] The indoor and outdoor temperature difference can be calculated by subtracting the outdoor temperature data from the indoor temperature data. The rate of change of humidity can be obtained by calculating the difference between the humidity of the current time period and the humidity of the previous time period, and then dividing it by the humidity of the previous time period. Whether it is in extreme weather conditions can be identified by obtaining meteorological data. When parameters such as temperature, humidity or wind speed exceed the preset extreme weather threshold, it is marked as extreme weather.

[0093] ④Equipment status characteristics:

[0094] Equipment operating time: reflects the cumulative operating time of the equipment and serves as an indicator for energy consumption and equipment maintenance.

[0095] Time since the last equipment maintenance: reflects the equipment maintenance history and helps predict equipment efficiency and energy consumption.

[0096] Identification of whether the equipment is under maintenance: used to mark whether the equipment is under maintenance so that the model can consider its impact on energy consumption.

[0097] The device operation time can be obtained by recording the time when the device started running and calculating the difference from the current time. The time since the last maintenance of the device can be obtained by recording the time when the device was last maintained and calculating the difference from the current time. Whether the device is in maintenance state can be identified by setting a maintenance state flag, which is set to true when the device is undergoing maintenance.

[0098] ⑤Time-related features:

[0099] Day of the week: reflects the difference between weekdays and weekends, which has an impact on energy consumption patterns.

[0100] Which day or month of the year: reflects the changes in seasons and months, and has a significant impact on energy consumption.

[0101] Holiday identification: used to mark holidays so that the model can consider their special impact on energy consumption.

[0102] ⑥Statistical characteristics:

[0103] Distribution characteristics of energy consumption data: describes the overall distribution of energy consumption data, which helps the model capture energy consumption patterns.

[0104] Rolling statistics of energy consumption data, such as rolling mean and rolling standard deviation, reflect the dynamic characteristics of energy consumption data.

[0105] The distribution characteristics of energy consumption data can be described by calculating the mean, median, mode, variance and other statistics of energy consumption data. The rolling statistics of energy consumption data can be obtained by calculating the rolling statistics of energy consumption data such as the mean and standard deviation over a period of time (such as the past hour).

[0106] ⑦ Change rate characteristics:

[0107] Energy consumption change rate: reflects the dynamic change of energy consumption over time, which is of great significance for predicting future energy consumption trends. The energy consumption change rate can be obtained by calculating the difference between the energy consumption of the current time period and the previous time period, and then dividing it by the energy consumption of the previous time period.

[0108] Embodiment 2:

[0109] The energy consumption prediction model is constructed based on the long short-term memory network (LSTM). The specific steps include:

[0110] First, the implementation steps include:

[0111] S301: Prepare a feature data set and divide it into a training set and a test set. The feature data set covers information in multiple dimensions, including time window features, human flow related features, environmental parameter features, equipment status features, time related features, statistical features, and change rate features, to ensure the comprehensiveness and accuracy of the energy consumption prediction model.

[0112] S302: Construct an LSTM network structure. The network structure includes an input layer, two hidden layers and an output layer. The input layer receives a feature data set as input. The two hidden layers contain 128 and 64 hidden units respectively, and each hidden unit updates and transmits information through a forget gate, an input gate and an output gate. The output of the hidden unit is nonlinearly transformed using a ReLU activation function to enhance the expressive power of the model. The output layer receives the output of the last hidden layer and maps the output of the LSTM network to the energy consumption prediction value through a fully connected layer.

[0113] S303: Train the LSTM model using the training set. Optimize the model parameters through the back propagation algorithm and the gradient descent method so that the model can accurately predict energy consumption. During the training process, the validation set is regularly used to evaluate the model performance to ensure the generalization ability of the model.

[0114] S304: Use the test set to evaluate the trained LSTM model. Verify the prediction performance of the model by comparing the predicted value of the model with the actual energy consumption value. The mean square error (MSE) is used as the evaluation index, and the algorithm formula is:

[0115]

[0116] Where n is the number of data points, T pred,i is the energy consumption value predicted by the model, Y actual,i is the actual measured energy consumption value; the prediction accuracy of the model is evaluated according to the MSE value. The smaller the MSE value, the higher the prediction accuracy of the model.

[0117] S305: Adjust the hyperparameters of the LSTM model according to the evaluation results. The hyperparameters include the number of network layers, the number of hidden layer units, the learning rate, etc. By adjusting the hyperparameters, the prediction accuracy and generalization ability of the model can be further improved.

[0118] S306: Train and save the corresponding LSTM energy consumption prediction model for each central air conditioner. Since the operating characteristics and environmental factors of each central air conditioner may be different, training the model separately can better meet the energy consumption prediction needs of different devices.

[0119] In the network structure of the LSTM model, the input layer is responsible for receiving the feature data set as input and passing the data to the hidden layer. The hidden layer is the core part of the LSTM network and contains two hidden layers, which are responsible for processing feature information at different levels. Each hidden unit updates and transmits information through the forget gate, input gate, and output gate, thereby capturing the long-term dependencies of time series data. The ReLU activation function is used to perform nonlinear transformation on the output of the hidden unit, enhancing the expressive power of the model. The output layer is responsible for mapping the output of the LSTM network to the energy consumption prediction value, and this mapping relationship is realized through the fully connected layer.

[0120] During the training process, the LSTM model is iteratively trained using the training set data. In each iteration, a batch of input data is fed into the LSTM model, and the values ​​of the input layer and the output layer are calculated to update the memory cells and hidden states, and finally the predicted output sequence is obtained. The model parameters are optimized by the back propagation algorithm and the gradient descent method, so that the model can gradually approach the real energy consumption data. At the same time, the validation set is regularly used to evaluate the model performance to ensure the generalization ability of the model. The learning rate is adjusted or the training is stopped early as needed to prevent overfitting.

[0121] The following is a code example of building and training an energy consumption prediction model using Python as the programming language:

[0122] import numpy as np

[0123] import pandas as pd

[0124] from keras.models import Sequential

[0125] from keras.layers import LSTM,Dense,ReLU

[0126] from sklearn.model_selection import train_test_split

[0127] from sklearn.preprocessing import StandardScaler

[0128] #DataFrame 'df' containing all features, and the 'target' column is the predicted energy consumption value

[0129] #Create a sample dataset

[0130] np.random.seed(0)

[0131] dates=pd.date_range('20230101', periods=100)

[0132] df=pd.DataFrame(np.random.randn(100,7),index=dates,columns=['feature1','feature2','feature3','feature4',

[0133] 'feature5','feature6','feature7'])

[0134] df['target'] = df.sum(axis = 1) + np.random.randn(100)*0.5 #Assumed energy consumption value

[0135] # Divide into training set and test set

[0136] X_train,X_test,y_train,y_test=train_test_split(features,targets,test_size=0.2,random_state=42)

[0137] #Build LSTM model

[0138] model=Sequential()

[0139] model.add(LSTM(128,return_sequences=True,input_shape=(1,X_train.shape[2])))

[0140] model.add(ReLU())

[0141] model.add(LSTM(64))

[0142] model.add(ReLU())

[0143] model.add(Dense(1))#output layer, 1 neuron corresponds to the energy consumption prediction value

[0144] # Compile the model

[0145] model.compile(optimizer='adam',loss='mean_squared_error')

[0146] #Train model

[0147] model.fit(X_train,y_train,epochs=50,batch_size=32,validation_split=0.1)

[0148] #Evaluate the model

[0149] mse=model.evaluate(X_test,y_test,verbose=0)

[0150] print(f'Test Mean Squared Error:{mse}')

[0151] #Save model

[0152] model.save('lstm_energy_prediction_model.h5')

[0153] The above code is an example of using Python programming language and Keras deep learning library to build and train energy consumption prediction model. The following is a detailed description of the code:

[0154] ①Import necessary libraries:

[0155] NumPy and Pandas are used for data processing.

[0156] keras.models.Sequential is used to build sequential models.

[0157] keras.layers.LSTM, keras.layers.Dense, keras.layers.ReLU layers used to build LSTM networks.

[0158] sklearn.model_selection.train_test_split is used to divide the training set and the test set.

[0159] sklearn.preprocessing.StandardScaler is used for feature standardization.

[0160] ② Create a sample data set: Use numpy to generate a DataFrame containing random numbers, simulate 7 features and a target energy consumption value. Use the date_range function of pandas to generate a date index.

[0161] ③ Divide the data set into training set and test set: Use the train_test_split function to divide the data set into training set and test set, and the test set size is 20%.

[0162] ④Build an LSTM model: Use the Sequential model to stack each layer sequentially.

[0163] Add two LSTM layers, the first with 128 hidden units and returning the sequence, and the second with 64 hidden units.

[0164] A ReLU activation function is added after each LSTM layer for nonlinear transformation.

[0165] Add a Dense layer as the output layer, which contains one neuron to predict the energy consumption value.

[0166] ⑤Compile model: Use adam optimizer and mean_squared_error loss function to compile the model.

[0167] ⑥Training model: Use the fit function to train the model, specify the training set, test set, number of iterations (epochs), batch size (batch_size), and validation set ratio.

[0168] ⑦Evaluate model: Use the evaluate function to evaluate the performance of the model on the test set and output the mean square error (MSE).

[0169] ⑧Save model: Use the save function to save the trained model to disk for subsequent use.

[0170] Embodiment 3:

[0171] In order to intuitively compare the real-time energy consumption data with the energy consumption data predicted by the model and use them for abnormal point detection, the present invention plots the real-time energy consumption data and the energy consumption data predicted by the model into curves that change over time, and aligns the timestamps of the two curves. The following is a detailed description and specific implementation of the method:

[0172] Obtain real-time energy consumption data from the energy consumption monitoring system or related data sources. These data usually contain timestamps and corresponding energy consumption values. At the same time, use the trained energy consumption prediction model, input the corresponding timestamp or time features, and obtain the energy consumption data predicted by the model.

[0173] In order to draw a curve that changes over time, it is necessary to ensure that the real-time energy consumption data and the energy consumption data predicted by the model are one-to-one in time. Therefore, timestamp alignment is a necessary step. Timestamp alignment can be achieved by following the steps below:

[0174] Determine the time range of the two data curves, namely the start time and the end time.

[0175] Select an appropriate time interval as the sampling interval of the curve as needed. For example, you can select every minute, every hour, or every day as the sampling interval.

[0176] For real-time energy consumption data, the energy consumption value of the corresponding timestamp is extracted according to the sampling interval. If a timestamp does not have a corresponding energy consumption value, interpolation processing can be performed to ensure that each timestamp has a corresponding energy consumption value.

[0177] For the energy consumption data predicted by the model, the predicted energy consumption value of the corresponding timestamp is also extracted according to the sampling interval. If there is no corresponding predicted value for a timestamp, the prediction model can be used for interpolation or extrapolation to obtain the predicted energy consumption value of the timestamp.

[0178] Using the timestamp as the horizontal axis and the energy consumption value as the vertical axis, the real-time energy consumption data curve and the energy consumption data curve predicted by the model are plotted respectively.

[0179] Make sure that the abscissas (time stamps) of the two curves are perfectly aligned to facilitate subsequent comparison and analysis.

[0180] For example, suppose that every hour is selected as the sampling interval and a 24-hour time range is determined. For real-time energy consumption data, the hourly energy consumption value is extracted from the energy consumption monitoring system and necessary interpolation processing is performed. For energy consumption data predicted by the model, the prediction model is used to input the hourly timestamp to obtain the corresponding predicted energy consumption value, and necessary interpolation or extrapolation processing is performed. Then, using the timestamp as the horizontal coordinate and the energy consumption value as the vertical coordinate, the real-time energy consumption data curve and the model-predicted energy consumption data curve are drawn in the coordinate system respectively, and the horizontal coordinates of the two curves are ensured to be completely aligned.

[0181] By plotting the real-time energy consumption data and the energy consumption data predicted by the model and aligning the timestamps, the change trends and differences of the two curves can be intuitively compared. This helps to further analyze the abnormal points of the energy consumption data and take corresponding measures for optimization and adjustment.

[0182] The present invention provides a variety of algorithms for marking abnormal points, aiming to effectively identify and mark abnormal points in energy consumption data. The following is a detailed description and specific implementation of each algorithm:

[0183] Difference threshold based algorithm:

[0184] This algorithm identifies abnormal points by setting a fixed energy consumption difference threshold T. In specific implementation, first obtain the real-time energy consumption data R and the predicted energy consumption data P, and then calculate the difference |RP| between the two. If the difference is greater than the preset threshold T, the corresponding real-time energy consumption data point is marked as an abnormal point.

[0185] For example, let T = 10. If the real-time energy consumption data R = 100 and the predicted energy consumption data P = 95 at a certain moment, then |RP| = 5, which is less than T, so the point is not marked as an abnormal point. If R = 110 and P = 95 at another moment, then |RP| = 15, which is greater than T, so the point is marked as an abnormal point.

[0186] Algorithm based on percentage of difference:

[0187] This algorithm identifies abnormal points by setting a percentage threshold PT of the energy consumption difference to the predicted value. In specific implementation, the percentage of the difference between the real-time energy consumption data R and the predicted energy consumption data P is calculated as |(RP) / P|. If the percentage is greater than the preset threshold PT, the corresponding real-time energy consumption data point is marked as an abnormal point.

[0188] For example, assuming PT = 0.1 (ie 10%), if R = 100 and P = 90 at a certain moment, then |(RP) / P| = |(100-90) / 90|≈0.11, which is greater than PT, so the point is marked as an outlier.

[0189] Algorithm based on sliding average:

[0190] This algorithm identifies abnormal points by calculating the average AVG of the energy consumption difference within a time window and setting a multiple M. In specific implementation, first calculate the average AVG of all energy consumption differences (RP) within the time window, and then determine whether the difference |RP| between the real-time energy consumption data and the predicted energy consumption data is greater than M times AVG. If so, the point is marked as an abnormal point.

[0191] For example, let the time window be the past 5 minutes, and M = 2. If the energy consumption differences in the window are -2, 3, -1, 4, and 2, respectively, then AVG = ((-2+3-1+4+2) / 5) = 1. If R = 100 and P = 97 at the current moment, then |RP| = 3, which is greater than 2 times AVG (i.e. 2), so the point is marked as an abnormal point.

[0192] Algorithms based on trend changes:

[0193] This algorithm identifies abnormal points by analyzing the changing trends of real-time energy consumption data and predicted energy consumption data. In specific implementation, the slopes Slope(R) and Slope(P) of the real-time energy consumption data curve and the predicted energy consumption data curve are first calculated, and then it is determined whether the difference between the two |Slope(R)-Slope(P)| is greater than the set slope difference threshold ST. If so, the point is marked as an abnormal point.

[0194] For example, let ST = 0.5. If, in a certain period of time, the linear regression calculation results in Slope(R) = 2, Slope(P) = 1, then |Slope(R)-Slope(P)| = 1, which is greater than ST, so the real-time energy consumption data point in this period is marked as an abnormal point.

[0195] An air conditioning energy consumption data monitoring system based on big data, the system comprising:

[0196] Data collection module: This module is responsible for collecting historical operating data from the central air-conditioning control system. These data cover the energy consumption data, operating settings, production parameters of each central air-conditioning unit, as well as environmental parameters and human flow records during operation. By connecting to the data interface of the central air-conditioning control system, the data collection module can obtain these data regularly or in real time and store them in the system's database for subsequent analysis and processing.

[0197] Feature construction module: After collecting the raw data, the feature construction module will preprocess and construct features for the data. This module will extract features that affect energy consumption prediction, such as the changing trend of environmental parameters, fluctuations in human traffic, and the operating status of equipment, and integrate these features into a feature data set. The formation of the feature data set provides the necessary data foundation for subsequent modeling.

[0198] Model building module: Using machine learning algorithms, the model building module builds energy consumption prediction models for each central air conditioner. These models are trained based on feature data sets, taking into account environmental parameters, changes in human traffic, equipment status, and time factors, to predict the energy consumption trend of central air conditioners over a period of time in the future. By continuously optimizing model parameters and algorithm selection, the accuracy and reliability of energy consumption prediction can be improved.

[0199] Real-time data acquisition module: This module is responsible for acquiring the current production parameters and operation settings of each central air conditioner, as well as the current environmental parameters and human flow data. Real-time data acquisition is achieved by connecting to the real-time data interface of the central air conditioning control system to ensure the timeliness and accuracy of the data.

[0200] Energy consumption prediction module: The energy consumption prediction module inputs real-time data into the energy consumption prediction model and receives the energy consumption prediction trend data output by the model. These prediction data can be used to compare with the real-time energy consumption data to detect energy consumption anomalies in a timely manner.

[0201] Data cleaning module: The data cleaning module collects the real-time energy consumption data of each central air conditioner from the central air conditioning control system and cleans the collected raw data. The cleaning process includes removing outliers and filling missing values ​​to ensure the accuracy and completeness of the data.

[0202] Anomaly detection module: The anomaly detection module plots the cleaned real-time energy consumption data and predicted energy consumption data into time-varying curves, and aligns the timestamps of the two curve data. The module determines whether each point on the curve has an anomaly based on a preset algorithm, and if so, marks it as an anomaly point. The specific choice of anomaly detection algorithm depends on the characteristics of the data and the needs of anomaly detection.

[0203] Warning signal generation and sending module: When the anomaly detection module finds an abnormal point, the warning signal generation and sending module will generate a warning signal according to the preset warning rules. The warning signal contains the specific information of the abnormal point, such as timestamp, real-time energy consumption value, predicted energy consumption value, abnormal type identification, and the number and location information of the relevant central air conditioner. The warning signal will be sent to the designated air conditioner management personnel through the system's communication interface so that they can take timely measures to deal with the abnormality.

[0204] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0205] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The air conditioning energy consumption data monitoring method based on big data is characterized by: The method comprises: S1: Collect historical operation data of the central air-conditioning control system, including energy consumption data, operation settings, production parameters of each central air-conditioning unit, as well as environmental parameters and human flow records during operation; S2: construct features for the data collected in step S1 to form a feature data set; S3: Using machine learning algorithms, an energy consumption prediction model is constructed for each central air conditioner. The energy consumption prediction model is based on a feature data set and comprehensively considers environmental parameters, changes in human traffic, equipment status, and time factors to predict the energy consumption trend of the central air conditioner in the future. S4: Obtain the current production parameters and operation settings of each central air conditioner, as well as the current environmental parameters and human flow data; S5: input the acquired data into the energy consumption prediction model, and receive the energy consumption prediction trend data output by the model; S6: Collect the real-time energy consumption data of each central air conditioner from the central air conditioning control system, and clean the collected raw data to remove abnormal values; S7: Plot the real-time energy consumption data collected in step S6 and the predicted energy consumption data output by the model in step S5 into curves that change over time, align the timestamps of the two curve data, and determine whether each point on the curve is abnormal according to a preset algorithm, and if so, mark it as an abnormal point; S8: Generate a warning signal based on the discovered abnormal points and warning rules, and further send the warning signal to the air conditioning management personnel.

2. The air conditioning energy consumption data monitoring method based on big data according to claim 1 is characterized in that: The feature data set includes: Time window characteristics, including average energy consumption, maximum energy consumption, and minimum energy consumption in the past preset time period; Crowd flow-related features, including the ratio of current crowd flow to the average crowd flow in a preset time period in the past, the rate of change of crowd flow, and the identification of crowd flow during peak hours; Environmental parameter characteristics, including indoor and outdoor temperature difference, humidity change rate, and whether it is in extreme weather conditions; Equipment status characteristics, including equipment operation time, time since the last equipment maintenance, and whether the equipment is under maintenance; Time-related features, including the day of the week, day or month of the year, and identification of holidays; Statistical characteristics, including the distribution characteristics of energy consumption data and the rolling statistics of energy consumption data; Rate of change characteristics, including the rate of change of energy consumption.

3. The air conditioning energy consumption data monitoring method based on big data according to claim 2 is characterized in that: The long short-term memory network LSTM algorithm is used to build an energy consumption prediction model. The specific steps include: S301: Using a feature data set as input, dividing the feature data set into a training set and a test set; S302: Build the LSTM network structure, set the number of network layers, the number of hidden layer units and the activation function hyperparameters; S303: Use the training set to train the LSTM model, and optimize the model parameters through the back propagation algorithm and the gradient descent method; S304: Use the test set to evaluate the trained LSTM model and verify the prediction performance of the model; S305: adjusting the hyperparameters of the LSTM model according to the evaluation results until a satisfactory prediction accuracy is achieved; S306: Train and save the corresponding LSTM energy consumption prediction model for each central air conditioner.

4. The air conditioning energy consumption data monitoring method based on big data according to claim 2 is characterized in that: The network structure of the LSTM model is: Input layer: receives feature data sets as input, which include time window features, crowd flow-related features, environmental parameter features, equipment status features, time-related features, statistical features, and change rate features; Network layer: contains two hidden layers, each containing 128 and 64 hidden units respectively. Each hidden unit updates and transmits information through the forget gate, input gate and output gate, and uses the ReLU activation function to perform nonlinear transformation on the output of the hidden unit. Output layer: The output layer receives the output of the last hidden layer and maps the output of the LSTM network to the energy consumption prediction value through a fully connected layer.

5. The air conditioning energy consumption data monitoring method based on big data according to claim 4 is characterized in that: The LSTM model is iteratively trained using the training set data. In each iteration, a batch of input data is fed into the LSTM model, and the values ​​of the input layer and the output layer are calculated, the memory cells and hidden states are updated, and finally the predicted output sequence is obtained; During training, periodically evaluate model performance using the validation set and adjust the learning rate or stop training early as needed to prevent overfitting.

6. The air conditioning energy consumption data monitoring method based on big data according to claim 5 is characterized in that: The mean square error (MSE) algorithm is used to evaluate the performance of the energy consumption prediction model. The specific algorithm is as follows: Where n is the number of data points, T pred,i is the energy consumption value predicted by the model, T actual,i is the actual measured energy consumption value; the prediction accuracy of the model is evaluated according to the MSE value. The smaller the MSE value, the higher the prediction accuracy of the model.

7. The air conditioning energy consumption data monitoring method based on big data according to claim 1 is characterized in that: The method of cleaning the original data in step S6 includes: setting a reasonable value range for each type of data, traversing the original data, marking the values ​​exceeding the set range as abnormal values, and removing the data marked as abnormal values ​​from the original data.

8. The air conditioning energy consumption data monitoring method based on big data according to claim 1 is characterized in that: The algorithms for marking outliers include: Difference threshold-based algorithm: Set a fixed energy consumption difference threshold. When the difference between the real-time energy consumption data and the predicted energy consumption data exceeds the threshold, it is marked as an abnormal point. The calculation formula is: if |RP|>T, it is marked as an abnormal point, where R represents the real-time energy consumption data, P represents the predicted energy consumption data, and T represents the set energy consumption difference threshold; Algorithm based on percentage of difference: set a percentage threshold of energy consumption difference to predicted value. When the percentage of difference between real-time energy consumption data and predicted energy consumption data to predicted value exceeds the threshold, it is marked as an abnormal point. The calculation formula is: if |(RP) / P|>PT, it is marked as an abnormal point, where R represents real-time energy consumption data, P represents predicted energy consumption data, and PT represents the set percentage threshold of energy consumption difference to predicted value. Algorithm based on sliding average: Calculate the average value of energy consumption difference within a time window. When the difference between real-time energy consumption data and predicted energy consumption data exceeds a certain multiple of the average value, it is marked as an abnormal point. The calculation formula is: if |RP|>M×AVG, it is marked as an abnormal point, where R represents real-time energy consumption data, P represents predicted energy consumption data, M represents the set multiple, and AVG represents the average value of energy consumption difference within the time window; Algorithm based on trend change: Analyze the changing trends of real-time energy consumption data and predicted energy consumption data. When the trend of real-time data deviates significantly from the trend of predicted data, it is marked as an abnormal point. The trend change is realized by calculating the slopes of the two curves. The formula for the slope difference is: if |Slope(R)-Slope(P)|>ST, it is marked as an abnormal point, where Slope(R) represents the slope of real-time energy consumption data, Slope(P) represents the slope of predicted energy consumption data, and ST represents the set slope difference threshold.

9. The air conditioning energy consumption data monitoring method based on big data according to claim 1 is characterized in that: The early warning signal includes the specific timestamp of the abnormal point, the corresponding real-time energy consumption value, the predicted energy consumption value, the abnormal type identification and the number and location information of the relevant central air conditioner.

10. An air conditioning energy consumption data monitoring system based on big data, characterized in that: The system comprises: The data collection module is responsible for collecting the historical operation data of the central air-conditioning control system, including the energy consumption data, operation settings, production parameters of each central air-conditioning unit, as well as the environmental parameters and human flow records during operation; The feature construction module constructs features of the collected data to form a feature data set for subsequent modeling; The model building module uses machine learning algorithms to build an energy consumption prediction model for each central air conditioner. Based on the feature data set, the model comprehensively considers environmental parameters, changes in human traffic, equipment status, and time factors to predict the energy consumption trend of the central air conditioner in the future. The real-time data acquisition module is responsible for obtaining the current production parameters and operation settings of each central air conditioner, as well as the current environmental parameters and human flow data; The energy consumption prediction module inputs real-time data into the energy consumption prediction model and receives the energy consumption prediction trend data output by the model; The data cleaning module collects the real-time energy consumption data of each central air conditioner from the central air conditioning control system and cleans the collected raw data to remove abnormal values; The anomaly detection module plots the cleaned real-time energy consumption data and predicted energy consumption data into time-varying curves, aligns the timestamps of the two curve data, and determines whether each point on the curve has an anomaly based on a preset algorithm. If so, it is marked as an abnormal point. The warning signal generation and sending module generates a warning signal based on the discovered abnormal points and the preset warning rules, and sends the warning signal to the designated air conditioning management personnel.

Images