Method and system for predicting aggregated load of central air conditioner

Abstract

The invention discloses a central air conditioner aggregation load prediction method and system, and relates to the technical field of central air conditioner control, and the method comprises the following steps: receiving the historical operation data of a central air conditioner, employing a Latin hypercube sampling LSH method to carry out the random and uniform sampling of the historical operation data of the central air conditioner, and obtaining the sample data of the central air conditioner, the historical operation data of the central air conditioner comprises temperature, humidity and electricity consumption; the central air conditioner sample data are input into a pre-established THBA-Bi-LSTM model, a central air conditioner aggregation load prediction result is obtained through output, and the pre-established THBA-Bi-LSTM model is obtained after parameters of Bi-LSTM are optimized through a THBA algorithm; the adjustable load of the central air-conditioning system can be managed and controlled more efficiently, and the purposes of energy conservation, stable operation and resource utilization optimization are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of central air-conditioning control, and in particular to a method and system for predicting aggregated load of a central air-conditioning. Background Art

[0002] With the rapid development of social economy, problems such as energy shortage and environmental pollution are becoming increasingly serious. The country is vigorously building a new energy system, and the proportion of renewable energy power generation is increasing. However, renewable energy power generation is unstable and intermittent. Large-scale access will bring power quality problems such as unstable frequency and severe harmonics to the power grid. At the same time, load diversification also makes the load side have great optimization potential, which will further enhance the capacity of the power grid after being introduced into the power grid. In addition, the power load of the power grid has increased year by year, and the proportion of temperature control loads such as central air conditioning in the power load remains high, and the contradiction between the power supply side and the demand side has become more and more obvious. Therefore, it is particularly important to promote high-precision prediction of central air-conditioning aggregate loads and optimize energy management to maintain the balance and stability of the power system.

[0003] In the field of central air conditioning control, there are currently some methods and systems for managing air conditioning load aggregation. For example, a method and device for evaluating the dynamic aggregation regulation potential of an air conditioning load cluster, that is, based on an equivalent thermal parameter model of the air conditioning load, according to the outdoor temperature, the upper limit of the indoor temperature setting value and the lower limit of the indoor temperature setting value, combined with the relationship between the air conditioning power and the room temperature change, the air conditioning load rotation control cycle is determined. In addition, a control strategy for a central air conditioner (CAC) load group to participate in the peak regulation of the power system is proposed, which proposes a central air conditioner single load model for participating in the peak regulation of the power system, and then constructs a system architecture for the central air conditioner load group to participate in the peak regulation of the power system, proposes an optimization model for the central air conditioner load group to participate in the peak regulation of the power system, and performs piecewise linearization on nonlinear constraints, so that the model can be solved quickly and effectively. Most of the existing technologies focus on the load control of a single adjustment parameter, and lack a method for accurately predicting the aggregated load of central air conditioners. Summary of the invention

[0004] In order to solve the deficiencies mentioned in the above background technology, the purpose of the present invention is to provide a central air-conditioning aggregate load prediction method and system, which can more efficiently manage and control the adjustable load of the central air-conditioning system.

[0005] In the first aspect, the purpose of the present invention can be achieved by the following technical solution: a method for predicting the aggregate load of a central air conditioner, the method comprising the following steps:

[0006] Receive historical operation data of the central air conditioner, and randomly and uniformly extract samples from the historical operation data of the central air conditioner using a Latin hypercube sampling (LSH) method to obtain central air conditioner sample data, wherein the historical operation data of the central air conditioner includes temperature, humidity, and power consumption;

[0007] The central air-conditioning sample data is input into a pre-established THBA-Bi-LSTM model, and the central air-conditioning aggregate load prediction result is output, wherein the pre-established THBA-Bi-LSTM model is obtained by optimizing the parameters of Bi-LSTM by using the THBA algorithm.

[0008] In combination with the first aspect, in certain implementations of the first aspect, the method further includes: the Latin hypercube sampling LSH method simulates the power variation law of each air-conditioning load and aggregates and superimposes the air-conditioning loads.

[0009] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of randomly and uniformly extracting samples from the historical operation data of the central air conditioner using the Latin hypercube sampling LSH method includes the following steps:

[0010] Let the random variables X1, X2, ..., X N The cumulative probability distribution function is Y N =F(X N ), first determine the sampling size N, and divide the sample into N parts evenly according to the probability distribution function;

[0011] Perform a random sampling in each interval, and in the i-th interval Y i =(r i -i-1) / N, where r i is a random number in [0,1];

[0012] Y i Substitute into Y N =F(X N ) is the inverse function F -1 , find the sampling value X of interval i i ;

[0013] Repeat the above steps N times to generate N variables that obey the F distribution function.

[0014] In combination with the first aspect, in some implementations of the first aspect, the method also includes: the THBA algorithm is obtained by introducing Tent chaotic mapping to improve the honey badger algorithm to initialize the population, and the THBA algorithm includes three stages, an initialization stage, a mining stage, and a honey collection stage.

[0015] In combination with the first aspect, in some implementations of the first aspect, the method further includes: a process in the initialization phase:

[0016] Set initialization parameters, including population size N, maximum number of iterations T max , using the Tent mapping formula (1) to generate the initial population of honey badgers with chaotic characteristics;

[0017]

[0018] Where, X i is the mapping function value at the i-th iteration; γ is the chaos coefficient; X i+1 is the mapping function value at the i+1th iteration;

[0019] Calculate the fitness value of the honey badger, obtain the optimal fitness value of the honey badger, and determine the food source;

[0020] The digging phase process: During the digging phase, the honey badger performs a heart-shaped movement, and the heart-shaped movement is simulated by formula (4):

[0021] Define the intensity I and update the density factor, where I is the odor intensity of the prey, as shown in formula (2):

[0022]

[0023] Where s is the source intensity or concentrated intensity; d i Represents the distance between the prey and the current honey badger individual, r2 is a random number in (0, 1);

[0024] The density factor is updated using the method in formula (3):

[0025]

[0026] Among them, T max is the maximum number of iterations; C ≥ 1;

[0027] Formula (4) simulates the action of the honey badger during the digging phase:

[0028] x new =x p +F×β×I×x p +F×r3×α×d i ×|cos(2πr4)×[1-cos(2πr5)]| (4)

[0029] Among them, x p is the global optimal position in the current state; β≥1 is the ability of the honey badger to obtain food, d i represents the distance between the prey and the current honey badger individual, r3, r4, r5 are three random numbers in (0, 1), and F is the sign of changing the search direction, which is determined using formula (5):

[0030]

[0031] The honey collection stage: the situation where the honey badger follows the honey guide bird to the beehive is simulated by formula (6):

[0032] x new =x p +F×r7×α×d i (6)

[0033] Among them, x new is the updated position of the honey badger individual; x p is the prey position; F and α are determined by formula (5) and formula (3) respectively; r7 is a random number in (0,1), obtained from formula (6), according to the distance information d i , honey badger in prey position x p Search nearby.

[0034] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of optimizing the parameters of the Bi-LSTM using the THBA algorithm includes:

[0035] The central air conditioning aggregate power data is used as the input feature variable, and the initial parameters to be optimized are set and given initial values; the parameters to be optimized are the learning rate, the number of neurons in the first hidden layer, the number of neurons in the second hidden layer, and the regularization coefficient;

[0036] During the training process, the root mean square error (MSE) between the predicted data and the actual data is used as the fitness function; the specific root mean square error (MSE) calculation formula is as follows:

[0037]

[0038] Among them, y t is the actual aggregate load value at time t, is the predicted aggregate load value at time t, N is the number of samples;

[0039] Run the improved THBA algorithm, perform search optimization Bi-LSTM, and adjust parameters to obtain the optimal solution.

[0040] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the Bi-LSTM includes a forget gate, an input gate, and an output gate, wherein a calculation formula of the forget gate is as follows:

[0041] f t =σ(W f [h t-1 ,x t ]+b f (8)

[0042] In the formula, W f is the weight matrix of the forget gate; b f is the bias variable of the forget gate, σ is the activation function sigmoid;

[0043] After the forget gate is the input gate, which also contains the weight matrix W. i And the bias value b of the input gate i , the calculation formula is as follows:

[0044] i t =σ(W i [h t-1 ,x t ]+b i (9)

[0045] The new cell state r of the current input t The calculation formula is as follows:

[0046] r t '=tanh(W r [h t-1 ,x t ]+b r (10)

[0047] The current unit state r t , the calculation method is based on the forget gate f at the current moment t and the unit state r at the previous moment t-1 After multiplication, the current unit state r t 'Multiply the current time input i t The result is added together, and the calculation formula is as follows:

[0048] r t =f t er t-1 +i t er t ' (11)

[0049] The calculation of the output gate of LSTM is shown in the following formula:

[0050] o t =σ(W o [h t-1 ,x t ]+b o ) (12)

[0051] The calculation expression of the output gate of LSTM, and the final output of LSTM is related to the output gate and the unit state. The calculation formula is as follows:

[0052] h t =ot e tanh(r t ) (13).

[0053] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the Bi-LSTM transmits the signal from the input layer to the network layer, obtains the output value through forward LSTM calculation, and obtains the output value through reverse LSTM calculation, and the calculation process is as follows:

[0054]

[0055] In the formula, x t is the input at time t, LSTM is the neural unit, is the output of the backward hidden layer at each moment, is the output of the hidden layer forward at each moment;

[0056] If the MSE reaches the minimum, the Bi-LSTM model is directly constructed with the parameters to predict the aggregated power of the central air conditioner, otherwise proceed to the next step;

[0057] If the MSE does not reach the minimum, it is determined whether the maximum number of iterations has been reached. If the maximum number of iterations has been reached, the obtained parameters are used to construct the THBA-Bi-LSTM model to realize the load prediction after central air conditioning aggregation.

[0058] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of inputting the central air conditioning sample data into a pre-established THBA-Bi-LSTM model and outputting a central air conditioning aggregate load prediction result includes:

[0059] Data preprocessing: Normalize the input data to the [0,1] interval to improve the training efficiency of the model. The normalization formula is as follows:

[0060]

[0061] Among them, x is the original data, x' is the normalized data, and x min is the minimum value in the original data, x max is the maximum value in the original data, which is the characteristic data related to the aggregate load.

[0062] Bi-LSTM model structure,specifically, input layer: receives input data, such as temperature, humidity, and power consumption.

[0063] Bi-LSTM layer: Contains forward LSTM layer and reverse LSTM layer, which process the forward and backward information of sequence data respectively. Among them, the forward LSTM layer calculates the forget gate, input gate, new cell state, cell state update and output gate, and the reverse LSTM layer is similar to the forward LSTM layer, but processes the reverse sequence of input data.

[0064] Aggregation output layer: Outputs the predicted aggregate load value, that is, merges the outputs of the forward and reverse LSTM layers, and outputs the final prediction result through a fully connected layer. The aggregate load prediction follows the following formula:

[0065]

[0066] Where W y is the output layer weight matrix, b y is the output layer bias vector, is the hidden state of the forward LSTM, is the hidden state of the backward LSTM, is the output value of aggregate load forecast at time t.

[0067] The calculation process of the forward LSTM layer and the reverse LSTM layer in the Bi-LSTM model is described in formulas (7) to (15).

[0068] In a second aspect, in order to achieve the above-mentioned object, the present invention discloses a central air conditioning aggregate load prediction system, comprising:

[0069] A data processing module is used to receive the historical operation data of the central air conditioner, and randomly and uniformly extract samples from the historical operation data of the central air conditioner using the Latin hypercube sampling LSH method to obtain central air conditioner sample data, wherein the historical operation data of the central air conditioner includes temperature, humidity and power consumption;

[0070] The load prediction module is used to input the central air-conditioning sample data into the pre-established THBA-Bi-LSTM model, and output the central air-conditioning aggregate load prediction result, wherein the pre-established THBA-Bi-LSTM model is obtained by optimizing the parameters of Bi-LSTM using the THBA algorithm.

[0071] Beneficial effects of the present invention:

[0072] Firstly, the present invention adopts Latin hypercube sampling LHS to perform central air conditioning aggregation modeling, which can realize the refined management of the adjustable load of the central air conditioning and improve the control accuracy and stability of the system. By improving the coverage and representativeness of the samples through LHS, the model is made more accurate and reliable, the system fluctuation is reduced, and the comfort of the indoor environment is improved.

[0073] Secondly, the adaptability of the central air-conditioning system in commercial scenarios is enhanced. Whether it is daily operation, peak power consumption or emergency demand response, the adjustable load can be quickly adjusted to meet the requirements of the system, improving the reliability and flexibility of the system.

[0074] Finally, the prediction model is improved by optimizing Bi-LSTM with the improved honey badger algorithm THBA, thereby realizing effective prediction of central air-conditioning load. Energy consumption is significantly reduced, and the goal of energy conservation and emission reduction is achieved. Accurately predict the load of central air-conditioning, reasonably adjust the load distribution, reduce unnecessary energy waste, and contribute to sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0076] Figure 1 It is a schematic flow chart of the method of the present invention;

[0077] Figure 2 This is a schematic diagram of the central air conditioning aggregate load prediction process based on the THBA-Bi-LSTM model of the present invention;

[0078] Figure 3 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0079] 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.

[0080] Embodiment 1:

[0081] The following is an introduction to the relevant terms involved in the embodiments of the present application:

[0082] Load forecasting: Load forecasting is to determine the load data at a specific moment in the future based on many factors such as the system's operating characteristics, capacity expansion decisions, natural conditions and social impacts, while meeting certain accuracy requirements. Load refers to the electricity demand (power) or electricity consumption. Load forecasting is an important part of the economic dispatch of power systems and an important module of the energy management system (EMS).

[0083] like Figure 1As shown, a method for predicting the aggregated load of a central air conditioner comprises the following steps:

[0084] S101: receiving historical operation data of a central air conditioner, and randomly and uniformly extracting samples from the historical operation data of the central air conditioner using a Latin hypercube sampling (LSH) method to obtain central air conditioner sample data, wherein the historical operation data of the central air conditioner includes temperature, humidity, and power consumption;

[0085] Among them, historical operating data of central air conditioners in commercial scenarios are collected and input, including but not limited to temperature, humidity, electricity consumption, etc.; the sampled variables and their distribution are determined based on the data characteristics.

[0086] In modern commercial buildings, in order to improve the operating efficiency and management level of central air-conditioning systems, efficient data acquisition technology has become particularly important. Although traditional data acquisition methods such as manual recording are low-cost, they have problems such as poor accuracy and high workload; while sensor networks and SCADA systems improve the accuracy and real-time performance of data, they have high initial investment and complex maintenance. Therefore, the present invention adopts the Internet of Things platform as the main means of central air-conditioning system operation data collection, aiming to achieve accurate monitoring and management of the system operation status in an automated manner.

[0087] The data collection method of the present invention first involves installing temperature and humidity sensors, electric energy meters and other equipment at various key locations of the central air conditioning system to ensure that they can accurately reflect the operating status of the system. Next, these devices are connected to the Internet of Things platform using wired or wireless communication technology to ensure the stability and reliability of data transmission. Then, a suitable data collection frequency is set, such as once every minute or hour, and the collected data is uploaded to a cloud server or a local data center for storage and management.

[0088] Furthermore, the present invention uses big data analysis and artificial intelligence technology to clean, process and analyze the collected data, generate visual reports, and help managers make more scientific decisions. This method not only achieves high-precision and real-time monitoring of data, but also reduces the need for manual intervention through automated data collection and management systems, thereby reducing operating costs. In addition, support for remote monitoring and control allows managers to view the system operation status at any time and discover and solve problems in a timely manner.

[0089] Finally, the data collection method of the Internet of Things platform provided by the present invention has good scalability and flexibility, and can easily add new sensors and devices according to needs. At the same time, the system maintenance is simple, and most problems can be solved through remote diagnosis and software updates. Modern Internet of Things platforms are usually equipped with advanced security mechanisms to ensure the security of data transmission and storage. In summary, the data collection method based on the Internet of Things platform is the best choice for central air-conditioning system operation data collection in commercial scenarios, which can effectively improve the system's operating efficiency and management level.

[0090] The Latin hypercube sampling LSH method simulates the power variation law of each air-conditioning load and aggregates and superimposes the air-conditioning load;

[0091] The process of randomly and uniformly extracting samples from the historical operation data of the central air conditioner using the Latin Hypercube Sampling (LSH) method comprises the following steps:

[0092] Assume that random variables X1, X2, ..., X N The cumulative probability distribution function is Y N =F(X N ). First, determine the sampling size N, and divide the sample into N parts evenly according to the probability distribution function;

[0093] Perform a random sampling in each interval, such as Y in the i-th interval i =(r i -i-1) / N, where r i is a random number in [0,1];

[0094] Y i Substitute into Y N =F(X N ) is the inverse function F -1 , find the sampling value X of interval i i ;

[0095] Repeat the above steps N times to generate N variables that obey the F distribution function.

[0096] The aggregate power characteristics of 3,000 commercial air conditioners are obtained, and the aggregate power data of central air conditioners in commercial scenarios is obtained.

[0097] S102: Input the central air-conditioning sample data into a pre-established THBA-Bi-LSTM model, and output a central air-conditioning aggregate load prediction result, wherein the pre-established THBA-Bi-LSTM model is obtained by optimizing the parameters of Bi-LSTM using the THBA algorithm.

[0098] The THBA algorithm is obtained by introducing the tent chaotic map to improve the honey badger algorithm to initialize the population. The THBA algorithm includes three stages: the initialization stage, the mining stage and the honey collection stage.

[0099] Initialization phase; set initialization parameters, such as population size N, maximum number of iterations T max etc., using the Tent mapping of formula (1) to generate the initial population of honey badgers with chaotic characteristics;

[0100]

[0101] Where, X i is the mapping function value at the i-th iteration; γ is the chaos coefficient; X i+1 is the mapping function value at the i+1th iteration.

[0102] Calculate the fitness value of honey badgers, obtain the optimal fitness value of honey badgers, and determine their food sources;

[0103] Digging phase: During the digging phase, the honey badger performs a heart-shaped motion. The heart-shaped motion can be simulated by formula (4):

[0104] First, define the intensity I and update the density factor. The intensity is related to the concentration of the prey and the distance between the prey and the honey badger. I is the smell intensity of the prey. The greater the smell intensity, the faster the honey badger moves, and vice versa. The specific formula is as shown in formula (2):

[0105]

[0106] Where s is the source intensity or concentrated intensity; d i Represents the distance between the prey and the current honey badger individual, and r2 is a random number in (0, 1).

[0107] The density factor ensures a smooth transition from exploration to exploitation. The decrease factor decreases with the number of iterations and is updated using the method in formula (3) to reduce the uncertainty caused by the time-varying changes in the simulated honey badger foraging process.

[0108]

[0109] Among them, T max is the maximum number of iterations; C≥1 (generally the default is 2).

[0110] Formula (4) simulates the action of the honey badger during the digging phase:

[0111] x new =x p +F×β×I×x p +F×r3×α×d i×|cos(2πr4)×[1-cos(2πr5)]| (4)

[0112] Among them, x p is the global optimal position in the current state; β≥1 (generally 6 by default) is the ability of the honey badger to obtain food, d i represents the distance between the prey and the current honey badger individual, r3, r4, r5 are three random numbers in (0, 1), and F is the sign of changing the search direction, which is determined using formula (5):

[0113]

[0114] During the excavation phase, honey badgers rely heavily on the intensity of the prey's scent, the distance to the prey, and the time-varying search influence factor α.

[0115] During the honey collection phase, the situation in which the honey badger follows the honey guide bird to the hive can be simulated using formula (6):

[0116] x new =x p +F×r7×α×d i (6)

[0117] Among them, x new is the updated position of the honey badger individual; x p is the prey position; F and α are determined by formula (5) and formula (3) respectively; r7 is a random number in (0,1). It can be seen from formula (6) that according to the distance information d i , honey badger in prey position x p Search nearby.

[0118] When selecting a prediction model, this application considered a variety of models, including traditional machine learning models such as support vector machines (SVM), random forests (RF), etc., as well as deep learning models such as convolutional neural networks (CNN), recurrent neural networks (RNN), etc. Finally, this application selected the Bi-LSTM model because it has significant advantages in processing time series data, can effectively capture time dependencies, maintain information integrity and reduce information loss. Bi-LSTM can utilize past and future information simultaneously through the combination of forward and reverse LSTM networks, significantly improving the accuracy of predictions.

[0119] In this application, we used the THBA algorithm to optimize the model by improving the honey badger initialization population using tent mapping. In order to prove the advantages of the THBA algorithm over other algorithms, we conducted the following comparative cases and experimental data analysis:

[0120] In the field of central air conditioning load forecasting, we used the THBA algorithm to optimize the bidirectional long short-term memory network (Bi-LSTM) and constructed the THBA-Bi-LSTM model. Through comparative experiments with GA-Bi-LSTM, PSO-Bi-LSTM and HBA-Bi-LSTM models, the THBA-Bi-LSTM model showed significant advantages in prediction accuracy. Specific experimental data show that the root mean square error (RMSE) of the THBA-Bi-LSTM model is 8.5 and the mean absolute error (MAE) is 5.2. These indicators are better than the GA-LSTM (RMSE is 11.6, MAE is 7.2), PSO-LSTM (RMSE is 10.5, MAE is 7..1) and HBA-Bi-LSTM (RMSE is 10.0, MAE is 7.0) models, proving the effectiveness and superiority of the THBA algorithm in optimizing the Bi-LSTM model.

[0121] The optimization capability of the THBA algorithm is due to its improved initialization population strategy through Tent mapping, which significantly improves the algorithm's global search capability and convergence speed. In the parameter optimization process of the Bi-LSTM model, the THBA algorithm can more effectively explore the solution space and find a better combination of parameters, thereby significantly improving the prediction accuracy of the model. In an experiment using the operating data of the central air-conditioning system in the past year, the coefficient of determination R 2 As for the prediction accuracy of the model, the prediction accuracy of the THBA-Bi-LSTM model on the test set reached 93%, while the prediction accuracy of the GA-Bi-LSTM, PSO-Bi-LSTM and HBA-Bi-LSTM models were 88%, 90% and 91% respectively. Therefore, the prediction accuracy of the THBA-Bi-LSTM model was improved by 5%, 3% and 2% respectively compared with the GA-LSTM and PSO-LSTM models. This shows that the THBA algorithm can more effectively explore the solution space and find a better parameter combination when optimizing the parameters of the Bi-LSTM model, thereby significantly improving the prediction accuracy of the model.

[0122] In practical applications, the high prediction accuracy of the THBA-Bi-LSTM model provides strong data support for the energy efficiency optimization of central air-conditioning systems. By accurately predicting load demand, the central air-conditioning system can be more effectively scheduled and controlled to achieve energy savings and cost reduction. These results not only verify the advantages of the THBA algorithm in theory, but also demonstrate its positive effects in the field of central air-conditioning load forecasting in practical applications, providing new optimization tools and methods for this field.

[0123] Therefore, by applying the THBA algorithm to the optimization of the Bi-LSTM model, we not only verified the wide applicability of the THBA algorithm in different fields, but also demonstrated its great potential in practical engineering applications. The high prediction accuracy of the THBA-Bi-LSTM model provides strong data support for the energy efficiency optimization of the central air-conditioning system. By accurately predicting the load demand, the central air-conditioning system can be more effectively dispatched and controlled to achieve energy savings and cost reduction. These results not only verify the advantages of the THBA algorithm in theory, but also demonstrate its positive effects in the field of central air-conditioning aggregate load prediction in practical applications.

[0124] In summary, through the analysis of comparative cases and experimental data, we can conclude that the THBA algorithm has obvious advantages over traditional GA, PSO and HBA algorithms in terms of global search capability, convergence speed and prediction effect improvement. These advantages have been proven not only in theory, but also in practical applications, providing new optimization tools for technicians in this field.

[0125] In summary, by selecting the Bi-LSTM model and the THBA optimization algorithm, this application not only performs well in processing time series data with long-term dependencies, but also shows stronger search capabilities and faster convergence speed in the optimization process, providing an efficient and accurate solution for the aggregate load prediction of the central air-conditioning system.

[0126] The process of optimizing the parameters of Bi-LSTM using the THBA algorithm includes:

[0127] The central air conditioning aggregate power data is used as the input feature variable, and the initial parameters to be optimized are set and given initial values; the parameters to be optimized are the learning rate, the number of neurons in the first hidden layer, the number of neurons in the second hidden layer, and the regularization coefficient;

[0128] Specifically, in HBA, the number of honey badgers is set to 20, the maximum number of iterations is 50, the search range of the learning rate of the optimized Bi-LSTM neural network is set to [0.001, 0.01], the search range of the number of nodes in the two hidden layers is [10, 80], the search range of the regularization coefficient is [0.0001, 0.001], and to prevent overfitting, the batch size is set to 64, the number of training times is set to 500, and the learning rate reduction factor is 0.1.

[0129] During the training process, the mean square error (MSE) between the predicted data and the actual data is used as the fitness function;

[0130] Run the improved THBA algorithm, perform search optimization Bi-LSTM, and adjust parameters to obtain the optimal solution; the specific root mean square error MSE calculation formula is as follows:

[0131]

[0132] Among them, y t is the actual aggregate load value at time t, is the predicted aggregate load value at time t, N is the number of samples;

[0133] The optimized Bi-LSTM model is calculated by the THBA-Bi-LSTM algorithm. Further, the specific steps of the Bi-LSTM algorithm include:

[0134] LSTM solves the long-term dependency problem of sequences by adding memory units to avoid gradient loss. Its key structure is three "gates", the forget gate, the input gate, and the output gate. The forget gate is used to select whether to retain or forget the cell state, the input gate selectively records new information in the cell state, and the output gate is used to transmit the control of the new cell state c at the current time t. t How much can be output and passed to LSTM as the output value h t The formulas for each part of LSTM are as follows:

[0135] The calculation formula of the forget gate is as follows:

[0136] f t =σ(W f [h t-1 ,x t ]+b f (8)

[0137] In the formula, W f is the weight matrix of the forget gate; b f is the bias variable of the forget gate, and σ is the activation function sigmoid.

[0138] After the forget gate is the input gate, which also contains the weight matrix W. i And the bias value b of the input gate i The calculation formula is as follows:

[0139] i t =σ(W i [h t-1 ,x t ]+b i (9)

[0140] After determining the forget gate and input gate, it is also necessary to describe the new unit state r of the current input t ′, the calculation formula is as follows:

[0141] r t '=tanh(W r [ht-1 ,x t ]+b r (10)

[0142] The current unit state r t , the calculation method is based on the forget gate f at the current moment t and the unit state r at the previous moment t-1 After multiplication, the current unit state r t 'Multiply the current time input i t The result is added together, and the calculation formula is as follows:

[0143] r t =f t er t-1 +i t er t ' (11)

[0144] The calculation of the output gate of LSTM is shown in the following formula:

[0145] o t =σ(W o [h t-1 ,x t ]+b o ) (12)

[0146] The calculation expression of the output gate of LSTM, and the final output of LSTM is related to both the output gate and the unit state. The calculation formula is as follows:

[0147] h t =o t e tanh(r t ) (13)

[0148] Bi-LSTM transfers the signal from the input layer to the network layer and obtains an output value through forward LSTM calculation, which will affect the learning rate and output function, etc. At the same time, the output value is obtained through reverse LSTM calculation, which will determine the connection weights of the input nodes in Bi-LSTM, thereby forming a new set of output values. The calculation process is as follows:

[0149]

[0150] In the formula, x t is the input at time t, LSTM is the neural unit, is the output of the backward hidden layer at each moment, is the output of the hidden layer at each moment forward.

[0151] If the MSE reaches the minimum, the Bi-LSTM model is directly constructed with the current parameters to predict the aggregated power of the central air conditioner, otherwise proceed to the next step;

[0152] If the MSE does not reach the minimum, it is determined whether the maximum number of iterations has been reached. If the maximum number of iterations has been reached, the obtained parameters are used to construct the THBA-Bi-LSTM model to realize the load prediction after central air conditioning aggregation.

[0153] The process of inputting the central air-conditioning sample data into the pre-established THBA-Bi-LSTM model and outputting the central air-conditioning aggregate load prediction result includes:

[0154] Data preprocessing: Normalize the input data to the [0,1] interval to improve the training efficiency of the model. The normalization formula is as follows:

[0155]

[0156] Among them, x is the original data, x' is the normalized data, and x min is the minimum value in the original data, x max is the maximum value in the original data, which is the characteristic data related to the aggregate load.

[0157] Bi-LSTM model structure,specifically, input layer: receives input data, such as temperature, humidity, and power consumption.

[0158] Bi-LSTM layer: Contains forward LSTM layer and reverse LSTM layer, which process the forward and backward information of sequence data respectively. Among them, the forward LSTM layer calculates the forget gate, input gate, new cell state, cell state update and output gate, and the reverse LSTM layer is similar to the forward LSTM layer, but processes the reverse sequence of input data.

[0159] Aggregation output layer: Outputs the predicted aggregate load value, that is, merges the outputs of the forward and reverse LSTM layers, and outputs the final prediction result through a fully connected layer. The aggregate load prediction follows the following formula:

[0160]

[0161] Where W y is the output layer weight matrix, b y is the output layer bias vector, is the hidden state of the forward LSTM, is the hidden state of the backward LSTM, is the output value of aggregate load forecast at time t.

[0162] The calculation process of the forward LSTM layer and the reverse LSTM layer in the Bi-LSTM model is described in formulas (7) to (15).

[0163] In order to verify the performance of the THBA-Bi-LSTM model, we conducted a detailed experimental comparison. The experiment used a year of central air-conditioning operation data of a commercial building, including temperature, humidity and electricity consumption. The selected comparison models include traditional Bi-LSTM, support vector machine (SVM), random forest (RF), GA-LSTM, PSO-LSTM and HBA-LSTM. The experimental results are shown in the following table:

[0164] Table 1 Prediction effect of aggregate load by different prediction methods

[0165]

[0166] As can be seen from the table above, the THBA-Bi-LSTM model outperforms other methods in all evaluation indicators. Specifically, compared with the traditional Bi-LSTM, the RMSE is reduced by 29.2%, the MAE is reduced by 30.7%, and the R 2 The improvement is 9.4%. This shows that the THBA-Bi-LSTM model has a significant improvement in prediction accuracy.

[0167] The THBA-Bi-LSTM method significantly outperforms other traditional methods in prediction performance. Specifically, compared with the traditional Bi-LSTM, the THBA-Bi-LSTM method reduces the RMSE by 29.2%, the MAE by 30.7%, and the R 2 Compared with SVM, RMSE is reduced by 43.3%, MAE is reduced by 42.2%, and R 2 Compared with random forest (RF), RMSE is reduced by 39.3%, MAE is reduced by 35.3%, and R 2 The improvement is 16.2%. These results fully demonstrate the significant advantages of the THBA-Bi-LSTM method in improving prediction accuracy and stability.

[0168] These data show that the THBA-Bi-LSTM model has significant advantages in central air-conditioning aggregate load forecasting and can provide more accurate forecasting results, thus helping to more effectively manage and control the load of the central air-conditioning system.

[0169] In order to verify the effectiveness of the THBA-Bi-LSTM model in the aggregate load forecasting of the central air-conditioning system, we conducted experiments on the central air-conditioning system of a commercial building. The system contains 3,000 air conditioners, and we collected one year of operating data, including hourly operating data such as temperature, humidity, and electricity consumption. By normalizing, feature extracting, and partitioning the data, we used Latin hypercube sampling to extract 10% of the data as a training set, and the rest as a test set to ensure the efficiency and generalization ability of model training.

[0170] In terms of experimental settings, we selected the THBA-Bi-LSTM model for aggregate load forecasting and compared it with the traditional LSTM model and the Bi-LSTM model. The evaluation indicators included the root mean square error (RMSE), mean absolute error (MAE), and determination coefficient (R 2 ). The experimental results show that the THBA-Bi-LSTM model outperforms the traditional model in all evaluation indicators, with RMSE of 8.5, MAE of 5.2, and R 2 It reaches 0.93, which shows that its prediction accuracy is significantly improved.

[0171] To further verify the practical application effect of the model, we selected three typical scenarios: high temperature in summer, low temperature in winter, and transitional seasons in spring and autumn for in-depth analysis. During the high temperature period in summer (July 1 to July 31), the RMSE of the model was 8.0, the MAE was 4.8, and the R 2 During the low temperature period in winter (January 1 to January 31), the RMSE is 9.0, the MAE is 5.6, and the R 2 is 0.92; in the spring and autumn transition seasons (March 1 to March 31 and October 1 to October 31), the RMSE is 8.5, the MAE is 5.2, and the R 2 These results consistently show that the THBA-Bi-LSTM model can maintain high prediction accuracy regardless of extreme weather conditions or normal conditions.

[0172] In summary, through detailed data collection, processing and comparative experiments, this study not only proves the superior performance of the THBA-Bi-LSTM model under different climatic conditions, but also provides strong technical support for the energy management and optimization of central air-conditioning systems, demonstrating its great potential in practical applications.

[0173] Specifically, the present invention is further described below by way of embodiments:

[0174] Embodiment 2: In the second aspect, as Figure 3 As shown, in order to achieve the above-mentioned purpose, the present invention discloses a central air conditioning aggregate load prediction system, comprising:

[0175] The data processing module 11 is used to receive the historical operation data of the central air conditioner, and randomly and uniformly extract samples from the historical operation data of the central air conditioner using the Latin hypercube sampling LSH method to obtain central air conditioner sample data, wherein the historical operation data of the central air conditioner includes temperature, humidity and power consumption;

[0176] The load prediction module 12 is used to input the central air-conditioning sample data into the pre-established THBA-Bi-LSTM model, and output the central air-conditioning aggregate load prediction result, wherein the pre-established THBA-Bi-LSTM model is obtained by optimizing the parameters of Bi-LSTM using the THBA algorithm.

[0177] Based on the same inventive concept, the present invention also provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the program instructions stored in the memory. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is used to implement one or more instructions, specifically for loading and executing one or more instructions in a computer storage medium to implement the above method.

[0178] It needs to be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium, on which a computer program is stored, and the computer program is executed by a processor to execute the above method. The storage medium can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0179] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0180] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure may have various changes and improvements, and these changes and improvements fall within the scope of the present disclosure to be protected.

Claims

1. A method for predicting the aggregate load of central air conditioners, characterized in that: The method comprises the following steps: Receive historical operation data of the central air conditioner, and randomly and uniformly extract samples from the historical operation data of the central air conditioner using a Latin hypercube sampling (LSH) method to obtain central air conditioner sample data, wherein the historical operation data of the central air conditioner includes temperature, humidity, and power consumption; The central air-conditioning sample data is input into a pre-established THBA-Bi-LSTM model, and the central air-conditioning aggregate load prediction result is output, wherein the pre-established THBA-Bi-LSTM model is obtained by optimizing the parameters of Bi-LSTM by using the THBA algorithm.

2. A method for predicting central air conditioning aggregate load according to claim 1, characterized in that: The Latin Hypercube Sampling (LSH) method simulates the power variation law of each air-conditioning load and aggregates and superimposes the air-conditioning loads.

3. A method for predicting central air conditioning aggregate load according to claim 1, characterized in that: The process of randomly and uniformly extracting samples from the historical operation data of the central air conditioner using the Latin Hypercube Sampling (LSH) method comprises the following steps: Let the random variables X1, X2, ..., X N The cumulative probability distribution function is Y N =F(X N ), first determine the sampling size N, and divide the sample into N parts evenly according to the probability distribution function; Perform a random sampling in each interval, and in the i-th interval Y i =(r i -i-1) / N, where r i is a random number in [0,1]; Y i Substitute into Y N =F(X N ) is the inverse function F -1 , find the sampling value X of interval i i ; Repeat the above steps N times to generate N variables that obey the F distribution function.

4. A method for predicting central air conditioning aggregate load according to claim 1, characterized in that: The THBA algorithm is obtained by introducing the tent chaotic map to improve the honey badger algorithm to initialize the population. The THBA algorithm includes three stages: the initialization stage, the mining stage and the honey collection stage.

5. A method for predicting central air conditioning aggregate load according to claim 4, characterized in that: The process of the initialization phase: Set initialization parameters, including population size N, maximum number of iterations T max , using the Tent mapping formula (1) to generate the initial population of honey badgers with chaotic characteristics; Where, X i is the mapping function value at the i-th iteration; γ is the chaos coefficient; X i+1 is the mapping function value at the i+1th iteration; Calculate the fitness value of the honey badger, obtain the optimal fitness value of the honey badger, and determine the food source; The digging phase process: During the digging phase, the honey badger performs a heart-shaped movement, and the heart-shaped movement is simulated by formula (4): Define the intensity I and update the density factor, where I is the odor intensity of the prey, as shown in formula (2): Where s is the source intensity or concentrated intensity; d i Represents the distance between the prey and the current honey badger individual, r2 is a random number in (0, 1); The density factor is updated using the method in formula (3): Among them, T max is the maximum number of iterations; C ≥ 1; Formula (4) simulates the action of the honey badger during the digging phase: x new =x p +F×β×I×x p +F×r3×α×d i ×|cos(2πr4)×[1-cos(2πr5)]| (4) Among them, x p is the global optimal position in the current state; β≥1 is the ability of the honey badger to obtain food, d i represents the distance between the prey and the current honey badger individual, r3, r4, r5 are three random numbers in (0, 1), and F is the sign of changing the search direction, which is determined using formula (5): The honey collection stage: the situation where the honey badger follows the honey guide bird to the beehive is simulated by formula (6): x new =x p +F×r7×α×d i (6) Among them, x new is the updated position of the honey badger individual; x p is the prey position; F and α are determined by formula (5) and formula (3) respectively; r7 is a random number in (0,1), obtained from formula (6), according to the distance information d i , honey badger in prey position x p Search nearby.

6. A method for predicting central air conditioning aggregate load according to claim 1, characterized in that: The process of optimizing the parameters of Bi-LSTM using the THBA algorithm includes: The central air conditioning aggregate power data is used as the input feature variable, and the initial parameters to be optimized are set and given initial values; the parameters to be optimized are the learning rate, the number of neurons in the first hidden layer, the number of neurons in the second hidden layer, and the regularization coefficient; During the training process, the root mean square error (MSE) between the predicted data and the actual data is used as the fitness function; the root mean square error (MSE) calculation formula is as follows: Among them, y t is the actual aggregate load value at time t, is the predicted aggregate load value at time t, N is the number of samples; Run the improved THBA algorithm, perform search optimization Bi-LSTM, and adjust parameters to obtain the optimal solution.

7. A method for predicting central air conditioning aggregate load according to claim 6, characterized in that: The Bi-LSTM includes a forget gate, an input gate and an output gate, wherein the calculation formula of the forget gate is as follows: f t =σ(W f [h t-1 ,x t ]+b f (8) In the formula, W f is the weight matrix of the forget gate; b f is the bias variable of the forget gate, σ is the activation function sigmoid; After the forget gate is the input gate, which also contains the weight matrix W. i And the bias value b of the input gate i , the calculation formula is as follows: i t =σ(W i [h t-1 ,x t ]+b i (9) The new cell state r' of the current input t The calculation formula is as follows: r' t =tanh(W r [h t-1 ,x t ]+b r (10) The current unit state r t , the calculation method is based on the forget gate f at the current moment t and the unit state r at the previous moment t-1 After multiplication, the current unit state r t 'Multiply the current time input i t The result is added together, and the calculation formula is as follows: r t =f t e r t-1 +i t e r' t (11) The calculation of the output gate of LSTM is shown in the following formula: the t =σ(W o [h t-1 ,x t ]+b o ) (12) The calculation expression of the output gate of LSTM, and the final output of LSTM is related to the output gate and the unit state. The calculation formula is as follows: h t =o t e fishy t ) (13)。 8. A method for predicting central air conditioning aggregate load according to claim 7, characterized in that: The Bi-LSTM is to transfer the signal from the input layer to the network layer, calculate the output value through the forward LSTM, and calculate the output value through the reverse LSTM. The calculation process is as follows: In the formula, x t is the input at time t, LSTM is the neural unit, is the output of the backward hidden layer at each moment, is the output of the hidden layer forward at each moment; If the MSE reaches the minimum, the Bi-LSTM model is directly constructed with the parameters to predict the aggregated power of the central air conditioner, otherwise proceed to the next step; If the MSE does not reach the minimum, it is determined whether the maximum number of iterations has been reached. If the maximum number of iterations has been reached, the obtained parameters are used to construct the THBA-Bi-LSTM model to realize the load prediction after central air conditioning aggregation.

9. A method for predicting central air conditioning aggregate load according to claim 1, characterized in that: The process of inputting the central air-conditioning sample data into the pre-established THBA-Bi-LSTM model and outputting the central air-conditioning aggregate load prediction result includes: Data preprocessing: Normalize the input data to the [0,1] interval. The normalization formula is as follows: Among them, x is the original data, x' is the normalized data, and x min is the minimum value in the original data, x max is the maximum value in the original data, wherein the original data is characteristic data related to the aggregated load; The Bi-LSTM model structure includes: Input layer: receiving input data, including temperature, humidity and power consumption; Bi-LSTM layer: It includes a forward LSTM layer and a reverse LSTM layer, which process the forward and backward information of the sequence data respectively. The forward LSTM layer calculates the forget gate, input gate, new cell state, cell state update and output gate, and the reverse LSTM layer processes the reverse sequence of the input data. Aggregation output layer: Outputs the predicted aggregate load value, that is, merges the outputs of the forward and reverse LSTM layers, and outputs the final prediction result through a fully connected layer. The aggregate load prediction follows the following formula: Where W y is the output layer weight matrix, b y is the output layer bias vector, is the hidden state of the forward LSTM, is the hidden state of the backward LSTM, is the output value of aggregate load forecast at time t.

10. A central air conditioning aggregate load forecasting system, characterized in that: include: A data processing module is used to receive the historical operation data of the central air conditioner, and randomly and uniformly extract samples from the historical operation data of the central air conditioner using the Latin hypercube sampling LSH method to obtain central air conditioner sample data, wherein the historical operation data of the central air conditioner includes temperature, humidity and power consumption; The load prediction module is used to input the central air-conditioning sample data into the pre-established THBA-Bi-LSTM model, and output the central air-conditioning aggregate load prediction result, wherein the pre-established THBA-Bi-LSTM model is obtained by optimizing the parameters of Bi-LSTM using the THBA algorithm.

Images