Central air conditioner chilled water system and energy-saving optimization control method

Abstract

The invention relates to a chilled water system of a central air conditioner and an energy-saving optimization control method. The energy-saving optimization method comprises the following steps: S1, establishing an energy consumption model of the chilled water system of the central air conditioner; s2, firstly performing correlation analysis on the selected cooling load influence factors by adopting a grey correlation analysis method, screening out strong correlation influence factors, and establishing a central air conditioner cooling load prediction model based on PSO-LSSVM (Particle Swarm Optimization-Least Square Support Vector Machine); s3, optimizing the energy consumption model parameters of the chilled water system of the central air conditioner while predicting the cooling load value of the next point by using an improved white whale algorithm, and enabling the output result of the energy consumption model to reach the minimum value; the system comprises an energy consumption modeling module, a cooling load prediction module and an energy-saving optimization module. According to the method, the energy consumption model and the cooling load prediction model are established, the cooling load value of the next point is predicted, parameter optimization is performed, and an optimization structure is output; the energy consumption model and the cold load prediction model can be updated periodically, and the problem of performance reduction caused by long-time equipment operation is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of central air-conditioning energy-saving optimization, and in particular to a central air-conditioning chilled water system and an energy-saving optimization control method. Background Art

[0002] In recent years, with the rapid development of urban construction in my country, the energy consumption level of buildings has increased significantly. According to statistics from relevant departments, my country's building energy consumption accounts for one-third of the country's total energy consumption. In the field of construction, the power consumption of central air conditioners is about 40%. Therefore, reducing building energy consumption and improving energy efficiency have become issues that need to be urgently addressed.

[0003] There are many ways to optimize energy conservation in building air conditioning at present, among which the four most core methods are reducing the demand for terminal cooling and heating loads, type selection optimization, system structure optimization and system operation parameter optimization. Among them, reducing the demand for cooling and heating loads is mainly achieved by improving the thermal insulation performance of the building maintenance structure, but the cost and maintenance cost are high. Type selection optimization is mainly to enhance the matching degree of each device in the system, such as avoiding the situation where the equipment operation is inefficient due to the over-selection of the chiller. For system structure optimization, variable air volume system and ice storage technology can be used. However, due to the complex structure of the air-conditioning system, there are problems such as difficult transformation and large investment amount for the three energy-saving methods of reducing the demand for cooling load, type selection optimization and system structure optimization. Therefore, optimizing the operating parameters of the air-conditioning system is the best way to reduce the energy consumption of the air-conditioning system.

[0004] The methods proposed so far are basically based on artificial intelligence, neural networks and other deep learning-based system energy-saving optimization, which will have the problems of large computational complexity, difficult on-site deployment, and easy to fall into local extreme values.

[0005] The invention patent with publication number CN1793746A discloses an automatic adjustment and control method for the flow of cooling water and chilled water in a central air-conditioning system. The central air-conditioning system includes a chiller, a cooling tower and an air-conditioning terminal connected to the chiller, a cooling water pump arranged on a cooling return water pipeline connecting the chiller and the cooling tower, a chilled water pump arranged on a chilled return water pipeline connecting the chiller and the air-conditioning terminal, and a PLC controller. A frequency converter connected to the LC controller is used to adjust the cooling water pump and the chilled water pump respectively, so that the flow of chilled water and cooling water in the central air-conditioning system can be accurately controlled, which can not only meet the use requirements of the central air-conditioning terminal but also achieve the purpose of energy saving, and the control accuracy is higher.

[0006] Therefore, it is an urgent problem to provide an energy-saving optimization method and system that can achieve energy-saving optimization, has low computational complexity and is not prone to falling into local extreme values. Summary of the invention

[0007] The purpose of the present invention is to provide a central air-conditioning chilled water system and an energy-saving optimization control method in order to overcome the defects of the above-mentioned prior art.

[0008] The purpose of the present invention can be achieved by the following technical solutions:

[0009] According to one aspect of the present invention, there is provided an energy-saving optimization control method for a central air-conditioning chilled water system, the energy-saving optimization method comprising the following process:

[0010] S1. Establish energy consumption model of central air-conditioning chilled water system;

[0011] S2. Firstly, the grey correlation analysis method is used to analyze the correlation of the selected cooling load influencing factors, and the strongly correlated influencing factors are screened out, and a central air-conditioning cooling load prediction model based on PSO-LSSVM is established;

[0012] S3. Use the improved Beluga algorithm to meet the prediction of the next point cooling load value, optimize the energy consumption model parameters of the central air-conditioning chilled water system, and make the output result of the energy consumption model reach the minimum value;

[0013] The energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model periodically update data, predict the cooling load through the real-time updated energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model, and optimize the energy consumption model at the same time.

[0014] As a preferred technical solution, the establishment of the energy consumption model includes the following process:

[0015] S11, collecting historical energy consumption data and equipment operation parameter data of the central air-conditioning chilled water system as data for energy consumption model parameter identification;

[0016] S12. Establish a mathematical model of equipment energy consumption of the central air-conditioning chilled water system, and use the least squares method to identify parameters based on the collected historical energy consumption data and equipment operation parameter data.

[0017] As an optimal technical solution, the historical energy consumption operation data includes the instantaneous power of the chiller and the instantaneous power of the chilled water pump; the equipment operation parameter data includes the cooling load of the chiller, the chilled water supply temperature, the chilled water return temperature, the cooling water return temperature, the chilled water pump operation frequency and the total chilled water flow.

[0018] As a preferred technical solution, the establishment of the central air-conditioning cooling load prediction model includes the following process:

[0019] S21, collecting central air-conditioning cooling load data and outdoor environment data as historical data for cooling load prediction model training;

[0020] S22, using the collected historical data as training samples for the cooling load prediction model, and using the particle swarm algorithm to perform penalty parameter and kernel function optimization training to obtain the final cooling load prediction model as the terminal constraint for the optimization of the central air-conditioning chilled water system;

[0021] S23, determining the prediction cycle, using the current cooling load data and the current outdoor environment data actually taken, as well as the cooling load data and the outdoor environment data of the previous point as inputs of the cooling load prediction model, and predicting the cooling load value of the next point.

[0022] As a preferred technical solution, the outdoor environment data includes outdoor meteorological data, and the outdoor meteorological data includes humidity and irradiance parameters.

[0023] As a preferred technical solution, the energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model update data include the following process:

[0024] S31, determine the optimization parameters of the central air-conditioning chilled water system, collect the real-time operation energy consumption data, equipment operation parameter data and outdoor environment data of the central air-conditioning chilled water system, and use the improved Beluga algorithm to solve the optimal system operation parameters while satisfying the predicted cooling load value of the next point;

[0025] S32, outputting the equipment operation parameter data of the final central air-conditioning chilled water system;

[0026] S33, determining whether the optimization cycle time of the central air-conditioning chilled water system has been reached, and repeating S31 to S33 until the optimization is stopped;

[0027] S34, determining the update time of the energy consumption model and the cooling load prediction model, and if the update time is reached, updating the energy consumption model parameters and the cooling load prediction model and storing them;

[0028] Among them, if the optimal cycle time of the central air-conditioning chilled water system is reached, the next point cooling load value, parameter optimization and output optimization structure are predicted.

[0029] As a preferred technical solution, the improvement of the White Whale algorithm includes the following process:

[0030] S101, initializing IBWO parameters, including the number of iterations and population size;

[0031] S102, initializing the population using a new population initialization strategy;

[0032] S103, calculating the fitness values ​​of individual beluga whales and selecting the individual with the best fitness;

[0033] S104. Calculate the balance factor B jand whale fall probability W j ;

[0034] S105, if B j If it is less than 0.5, it is in the development stage, otherwise it is in the exploration stage;

[0035] S106, if B j Greater than W j , it is the whale fall stage, otherwise it is the reverse learning strategy;

[0036] S107, updating fitness to determine the best individual;

[0037] S108. If the maximum number of iterations is reached, the parameter adjustment result is output, otherwise S103 is repeated until the maximum number of iterations is reached.

[0038] According to another aspect of the present invention, a central air-conditioning chilled water system adopting any one of the above-mentioned systems comprises an energy consumption modeling module, a cooling load prediction module and an energy-saving optimization module;

[0039] Among them, the energy consumption modeling module establishes the energy consumption model of the chiller and the energy consumption model of the chilled water pump, and the cooling load prediction module establishes the cooling load prediction model; the energy consumption modeling module and the cooling load prediction module transmit data to the energy-saving optimization module, and the energy-saving optimization module processes the received data and outputs the optimization parameters.

[0040] As a preferred technical solution, the energy-saving optimization module includes an objective function unit, an optimization algorithm unit and a constraint condition unit. The data of the energy consumption modeling module is transmitted to the objective function unit, and the data of the cooling load prediction module is transmitted to the constraint condition unit. The objective function unit and the constraint condition unit output optimization parameters through the improved Beluga algorithm in the optimization algorithm unit.

[0041] As a preferred technical solution, the energy-saving optimization module also includes chiller equipment data, refrigerated water pump equipment data, equipment physical constraints and constraints between equipment. The chiller equipment data and refrigerated water pump equipment data are transmitted to the objective function unit, and the equipment physical constraints and constraints between equipment are transmitted to the constraint condition unit.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The present invention predicts the next point cooling load value by establishing an energy consumption model and a cooling load prediction model, and performs parameter optimization and output optimization structure.

[0044] 2. Before establishing the cooling load prediction model, the present invention uses the grey correlation analysis method to perform correlation analysis on the relevant influencing factors of the cooling load, selects the strongly correlated influencing factors, reduces the training amount for the training of the cooling load model, and shortens the training time.

[0045] 3. The present invention periodically updates the energy consumption model and the cooling load prediction model, thereby avoiding the problem of performance degradation caused by long-term equipment operation, which results in the model parameters not being consistent with the site.

[0046] 4. The present invention adopts the improved Beluga algorithm to optimize the parameters of the central air-conditioning chilled water system, which reduces the training time of the model to a certain extent, and avoids the problem of falling into local extreme values ​​when using the neural network algorithm. The improved Beluga algorithm adopts a new population initialization strategy and reverse learning strategy, which improves the convergence speed and accuracy of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a framework diagram of the energy-saving optimization method of the present invention;

[0048] Figure 2 This is a flow chart of an example of the energy-saving optimization method of the present invention;

[0049] Figure 3 The flowchart of the improved Beluga algorithm of the present invention is as follows;

[0050] Figure 4 is the cooling load prediction result of the present invention;

[0051] Figure 5 This is a comparison diagram before and after the energy consumption optimization of the present invention;

[0052] Figure 6 This is an energy-saving effect diagram of the present invention. DETAILED DESCRIPTION

[0053] 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 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 should fall within the scope of protection of the present invention.

[0054] In recent years, with the rapid development of urban construction in my country, the energy consumption level of buildings has increased significantly. According to statistics from relevant departments, my country's building energy consumption accounts for one-third of the country's total energy consumption. In the field of construction, the power consumption of central air conditioners is about 40%. Therefore, reducing building energy consumption and improving energy efficiency have become issues that need to be urgently addressed.

[0055] There are many ways to optimize energy conservation in building air conditioning at present, among which the four most core methods are reducing the demand for terminal cooling and heating loads, type selection optimization, system structure optimization and system operation parameter optimization. Among them, reducing the demand for cooling and heating loads is mainly achieved by improving the thermal insulation performance of the building maintenance structure, but the cost and maintenance cost are high. Type selection optimization is mainly to enhance the matching degree of each device in the system, such as avoiding the situation where the equipment operation is inefficient due to the over-selection of the chiller. For system structure optimization, variable air volume system and ice storage technology can be used. However, due to the complex structure of the air-conditioning system, there are problems such as difficult transformation and large investment amount for the three energy-saving methods of reducing the demand for cooling load, type selection optimization and system structure optimization. Therefore, optimizing the operating parameters of the air-conditioning system is the best way to reduce the energy consumption of the air-conditioning system.

[0056] The technical concept of the present invention is as follows: based on the adopted equipment energy consumption data and equipment operating parameters, least squares method parameter identification is performed to obtain an accurate energy consumption model of the equipment; combining the cooling load data and meteorological data, more specifically, using the grey correlation analysis method to screen out data with a greater correlation with the cooling load, and providing core data for the cooling load prediction model; obtaining the cooling load prediction model by training the selected cooling load prediction model and saving it, and predicting the cooling load value of the next point according to the on-site cooling load, meteorological and other data; finally, using the improved Beluga algorithm to dynamically adjust the equipment operating parameters while meeting the cooling load demand, thereby achieving the purpose of energy saving.

[0057] The present invention provides a central air-conditioning chilled water system and an energy-saving optimization control method; the present invention predicts the next point cold load value by establishing an energy consumption model and a cold load model, performs parameter optimization and outputs an optimization structure; before establishing the cold load prediction model, the present invention adopts a grey correlation analysis method to perform correlation analysis on relevant influencing factors of the cold load, selects strongly correlated influencing factors, reduces the training amount for the training of the cold load model, and shortens the training time; the present invention periodically updates the energy consumption model and the cold load prediction model, avoids the situation that the performance is degraded due to long-term equipment operation, and causes the model parameters to be inconsistent with the site; the present invention adopts an improved Beluga algorithm to optimize the parameters of the central air-conditioning chilled water system, reduces the training time of the model to a certain extent, and also avoids the problem of falling into a local extreme value by adopting a neural network algorithm, and adopts a new population initialization strategy and reverse learning strategy by adopting the improved Beluga algorithm, thereby improving the convergence speed and accuracy of the algorithm.

[0058] Example 1

[0059] like Figure 1 As shown, an energy-saving optimization control method for a central air-conditioning chilled water system is characterized in that the energy-saving optimization method includes the following processes:

[0060] S1. Establish energy consumption model of central air-conditioning chilled water system;

[0061] S2. Firstly, the grey correlation analysis method is used to analyze the correlation of the selected cooling load influencing factors, and the strongly correlated influencing factors are screened out, and a central air-conditioning cooling load prediction model based on PSO-LSSVM is established;

[0062] S3. Use the improved Beluga algorithm to meet the prediction of the next point cooling load value, optimize the energy consumption model parameters of the central air-conditioning chilled water system, and make the output result of the energy consumption model reach the minimum value;

[0063] The energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model periodically update data, predict the cooling load through the real-time updated energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model, and optimize the energy consumption model at the same time.

[0064] In this embodiment, the energy consumption model includes a refrigerator energy consumption model and a refrigeration pump energy consumption model. According to statistics from a large number of literature, the energy consumption models of refrigerators and refrigeration water pumps are as follows:

[0065] P ch =a+b(T co -T wo )+c(T co -T wo ) 2 +dQ+eQ 2 +f(T co -T wo )Q

[0066] Where P ch is the energy consumption of the chiller (kW); Q is the cooling capacity of the chiller (kW); T co is the cooling tower outlet water temperature (℃); T wo is the chilled water supply temperature (°C). a, b, c, d, e, and f are the parameters that need to be solved.

[0067] P chiled pump =b1+b2·L+b3·L 2 +b4·L 3

[0068] Where P chiled pump is the power consumption of the chilled water pump; b1, b2, b3, b4 are the performance coefficients of the chilled water pump; L is the flow rate of the chilled water pump.

[0069] According to the similarity of the water pumps:

[0070]

[0071] Where, L chiled is the rated flow of the chilled water pump; n is the actual speed of the water pump; n chiled is the rated speed of the chilled water pump; w is the speed ratio of the chilled water pump.

[0072] Before the prediction model is established, the grey correlation analysis method is used to conduct correlation analysis on the selected cooling load influencing factors to screen out the strongly correlated influencing factors. The principle of the grey correlation analysis method is as follows:

[0073] Grey Relational Analysis (GRA) is a grey system analysis method based on "small sample" and "poor information". Grey Relational Analysis is applicable to any regular or irregular samples, and the amount of calculation is small during analysis, so it is widely used. The specific steps are as follows:

[0074] First, determine the evaluation object and influencing factors, and establish an evaluation matrix. Assuming there are m influencing factors (i = 1, 2, L, m), and each influencing factor has n data (k = 1, 2, L, n), the evaluation matrix can be established as:

[0075]

[0076] Step 1: Determine the reference sequence and comparison sequence, and record the cooling load data column as the reference sequence X j (k), k = 1, 2, L, n, j = 1, 2, L, h, where n represents the length of the cooling load data column and h represents the number of cooling loads. The cooling load influencing factor data column is recorded as the comparison sequence X i (k), k = 1, 2, L, n, i = 1, 2, L, m, where n is the length of the influencing factor data column and m represents the number of influencing factors.

[0077] Step 2: Dimensionless processing of the data. Since the physical meanings and measurement units of the data of various influencing factors are different, different dimensions will be generated, resulting in a large difference in order of magnitude. Therefore, the data should be dimensionless (averaged):

[0078]

[0079] Where, X j (k) is the kth data of the jth reference sequence; is the average value of the jth reference sequence; X i (k) is the kth data of the i-th comparison sequence; is the average value of the ith comparison sequence.

[0080] Step 3: Calculate the absolute value of the difference between the reference sequence and the comparison sequence.

[0081] Δ ji (k) = |X' j (k)-X i '(k)|

[0082] Where X' j (k) and X i '(k) represent the jth dimensionless reference sequence and the ith dimensionless comparison sequence respectively.

[0083] Step 4: Find the secondary Δ max and Δ min :

[0084] Δ max =max i max k Δ ji (k)

[0085] Δ min =min i min k Δ ji (k)

[0086] In the formula, Δ max and Δ min Represent the maximum and minimum absolute differences of the second level respectively.

[0087] Step 5: Calculate the correlation coefficient ξ of each influencing factor i (k):

[0088] ξ i (k)=(Δ min +ρΔ max ) / Δ ji (k)+ρΔ max

[0089] In the formula, ρ is the resolution coefficient, where ρ∈[0,1], and ρ=0.5 is generally taken.

[0090] Step 6: Calculate each influencing factor X i (k) Relative cooling load X j The entropy weighted correlation degree r of (k) i :

[0091]

[0092] The establishment of the energy consumption model includes the following processes:

[0093] S11, collecting historical energy consumption data and equipment operation parameter data of the central air-conditioning chilled water system as data for energy consumption model parameter identification;

[0094] S12. Establish a mathematical model of equipment energy consumption of the central air-conditioning chilled water system, and use the least squares method to identify parameters based on the collected historical energy consumption data and equipment operation parameter data.

[0095] In this embodiment, the principle of the least squares support vector machine is as follows:

[0096] The least squares support vector machine is a machine learning algorithm based on the SVM, that is, the improved SVM. When the least squares support vector machine handles regression problems such as load forecasting, it uses a nonlinear mapping function The original nonlinear problem in the low-dimensional space is transformed into a linear regression in the high-dimensional feature space, that is, the input and output of the sample are fitted in the high-dimensional space:

[0097]

[0098] Where w is the weight vector and b is the bias.

[0099] The structural risk minimization principle is used when defining the optimization problem, as shown in the following formula:

[0100]

[0101] Where: C is the penalty parameter. The higher the value of C, the more likely it is to overfit. On the contrary, the lower the value, the more likely it is to underfit. Therefore, it is necessary to make a reasonable choice of C. i is the fitting error.

[0102] In order to solve the above optimization problem, the Lagrangian function is used to solve the optimization problem:

[0103]

[0104] In the formula, α i ∈R represents the Lagrange multiplier.

[0105] According to the KKT conditions, the above formula is optimized, that is, w, b, e i and α i The partial derivative of is equal to 0, so:

[0106]

[0107] Thus, after eliminating variables w and e i After that, the above optimization problem becomes a problem of solving linear equations, as shown in the following formula:

[0108]

[0109] In the formula, α=[α1,α2,L,α l], Q=[1,1,…,1], Y=[y1,y2,…,y l ] T , I represents the identity matrix, K(x i ,x j ) represents the selected kernel function, and defines the kernel function If the necessary and sufficient condition of the kernel function, Mercer's principle, is satisfied, the prediction model can be expressed as:

[0110]

[0111] In the formula, α i and b can be obtained by solving the above equation.

[0112] The kernel function uses the radial basis kernel function, and the formula is as follows:

[0113]

[0114] The prediction performance of the LSSVM model is determined by the penalty parameter C and the kernel parameter σ. The PSO algorithm is used to optimize these two parameters. The PSO algorithm is a more traditional algorithm.

[0115] The historical energy consumption operation data includes the instantaneous power of the chiller and the instantaneous power of the chilled water pump; the equipment operation parameter data includes the cooling load of the chiller, the chilled water supply temperature, the chilled water return temperature, the cooling water return temperature, the chilled water pump operation frequency and the total chilled water flow.

[0116] The establishment of the central air conditioning cooling load prediction model includes the following processes:

[0117] S21, collecting central air-conditioning cooling load data and outdoor environment data as historical data for cooling load prediction model training;

[0118] S22, using the collected historical data as training samples for the cooling load prediction model, and using the particle swarm algorithm to perform penalty parameter and kernel function optimization training to obtain the final cooling load prediction model as the terminal constraint for the optimization of the central air-conditioning chilled water system;

[0119] S23, determining the prediction cycle, using the current cooling load data and the current outdoor environment data actually taken, as well as the cooling load data and the outdoor environment data of the previous point as inputs of the cooling load prediction model, and predicting the cooling load value of the next point.

[0120] The outdoor environment data includes outdoor meteorological data, and the outdoor meteorological data includes humidity and irradiance parameters.

[0121] like Figure 2As shown, the energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model update data include the following process:

[0122] S31, determine the optimization parameters of the central air-conditioning chilled water system, collect the real-time operation energy consumption data, equipment operation parameter data and outdoor environment data of the central air-conditioning chilled water system, and use the improved Beluga algorithm to solve the optimal system operation parameters while satisfying the predicted cooling load value of the next point;

[0123] S32, outputting the equipment operation parameter data of the final central air-conditioning chilled water system;

[0124] S33, determining whether the optimization cycle time of the central air-conditioning chilled water system has been reached, and repeating S31 to S33 until the optimization is stopped;

[0125] S34, determining the update time of the energy consumption model and the cooling load prediction model, and if the update time is reached, updating the energy consumption model parameters and the cooling load prediction model and storing them;

[0126] Among them, if the optimal cycle time of the central air-conditioning chilled water system is reached, the next point cooling load value, parameter optimization and output optimization structure are predicted.

[0127] like Figure 3 As shown, the improvement of the White Whale algorithm includes the following processes:

[0128] S101, initializing IBWO parameters, including the number of iterations, population size and other IBWO parameters;

[0129] S102, initializing the population using a new population initialization strategy;

[0130] S103, calculating the fitness values ​​of individual beluga whales and selecting the individual with the best fitness;

[0131] S104. Calculate the balance factor B j and whale fall probability W j ;

[0132] S105, if B j If it is less than 0.5, it is in the development stage, otherwise it is in the exploration stage;

[0133] S106, if B j Greater than W j , it is the whale fall stage, otherwise it is the reverse learning strategy;

[0134] S107, updating fitness to determine the best individual;

[0135] S108. If the maximum number of iterations is reached, the parameter adjustment result is output; otherwise, S103 is repeated until the maximum number of iterations is reached.

[0136] In this embodiment, the basic white whale optimization algorithm (BWO) is an optimization algorithm proposed to simulate the hunting and survival behaviors of white whales. [6] , which can help solve various optimization problems.

[0137] The essence of BWO is to use white whales as search agents, and each white whale corresponds to a candidate solution. In BWO, the problem of each agent is modeled in the form of a matrix, which covers the position information of each white whale. The code position matrix of BWO is usually a two-dimensional matrix, where each row represents a search agent (white whale) and each column represents a different dimension or feature of the search agent. The position matrix of the search agent can be expressed as:

[0138]

[0139] In the above formula, n is the number of beluga whales; d is the dimension. The fitness value F of the beluga whale population x for:

[0140]

[0141] Balance Factor B f The calculation formula is shown below, which is used to determine whether Beluga enters the exploration or development stage.

[0142]

[0143] Where: T represents the current number of iterations; T max is the maximum number of iterations; B f The range is between 0 and 1. f When it is less than 0.5, the algorithm enters the development stage; when B f is greater than 0.5, the algorithm enters the exploration phase. As T increases, B f The fluctuation range decreases from (0,1) to (0,0.5), and the probability in the development stage will continue to increase.

[0144] The algorithm enters the exploration phase, and the update position formula of the beluga population is as follows:

[0145]

[0146] In the above formula: T is the current number of iterations; is the new position of the i-th beluga whale in the j-th dimension; p j is a random integer selected within the range of dimension j; For the i-th beluga whale at p j The location of the dimension; and are the current positions of the rth beluga whale and the ith beluga whale respectively; r is a randomly selected beluga whale; r1∈(0,1); r2∈(0,1).

[0147] BWO introduced the Levy flight strategy during the development phase, which is conducive to improving the convergence of the algorithm. The calculation method is:

[0148]

[0149] Where: T is the current iteration number; is the location of a randomly selected beluga whale; The location of the i-th beluga whale; is the position of the ith beluga whale at the T+1th iteration; r3 and r3 are random numbers between 0 and 1; C1 represents the random jump intensity of Levy flight; the expression of C1 is:

[0150] C1=2r4(1-r / r max )

[0151] L F To meet the random number of Levy flight, the expression is:

[0152]

[0153] u and v are random numbers that satisfy the normal distribution, β=0.5.

[0154] The BWO algorithm introduces a simulated whale fall behavior to simulate randomness and changes in the beluga group. The calculation formula is as follows:

[0155]

[0156] Where: r5, r6 and r7 are random numbers between 0 and 1; X step is the step length of the whale's fall. step The expression is:

[0157] X step =(u b -l b )exp(-C2T / T max )

[0158] Where: u b and l b are the upper and lower limits of the variable respectively; C2 is the step factor related to the size of the population and the probability of whale decline. The expression of C2 is:

[0159] C2=2W f ×n

[0160] Where: W fThe probability of a whale falling.

[0161] W f It is a linear function, which gradually decreases from 0.1 at the beginning to 0.05. Its calculation expression is:

[0162]

[0163] The improvement is to adopt an improved new population initialization strategy and reverse learning strategy:

[0164] The traditional BWO algorithm population initialization adopts a random distribution method, which will lead to poor quality of the initial solution in the early stage and easily fall into the local optimum. Therefore, to solve this problem, this paper proposes an initialization method that combines Tent mapping and Sobol mapping. This method combines the advantages of both, which can make the initial solution distribution of the BWO algorithm more uniform and expand the coverage. The calculation formula of the proposed new initialization strategy is as follows:

[0165]

[0166] As an important mathematical concept, reverse learning is mainly used to enhance the overall search ability of the algorithm. [9] . Reverse learning can provide an opposite orientation for the individual position of the beluga whale, thereby exploring another possibility, expanding the search range, and avoiding falling into local extremes. Finally, the two solutions are compared, and the current optimal solution is used as the position of the individual beluga whale. At present, the directional learning strategy has been applied in many optimization algorithms. The mathematical expression of reverse learning is:

[0167] X i '=(ub+lb)×β-X i

[0168] ub and lb are the upper and lower limits of the search space, respectively, and β is a random number between 0:1.

[0169] The least square method is used to identify the parameters of the energy consumption model, determine the input parameters and output parameters of the chilled water system, and form and store the energy consumption model of each device during the debugging stage.

[0170] The optimized PSO-LSSVM regression prediction algorithm is used to train the model based on outdoor meteorological data and cooling load data and use them as historical data. This method has short training time, is easy to deploy, and can accurately predict future cooling load demand.

[0171] Preferably, the LSSVM model before optimization is a machine learning algorithm improved on the basis of SVM, and the kernel function used is the radial basis kernel function. Since there are two preliminary steps in the algorithm that need to be set by human experience and have a certain degree of blindness, PSO is used to optimize and select these two parameters.

[0172] Energy-saving optimization of central air-conditioning chilled water system realized by improved Beluga algorithm. The optimization algorithm adopts the predicted value of cooling load as the equality constraint; the chilled water system includes chilled water main unit and chilled water pump, and the energy consumption model parameter identification of the equipment adopts the least square method.

[0173] The air conditioning energy-saving optimization method adopts an improved Beluga algorithm, and the parameter settings of the improved Beluga algorithm are:

[0174] Optimization objective function: The sum of the energy consumption of all equipment in the chilled water system of the air-conditioning system is taken as the optimization target of the system, so that the total energy consumption of the equipment is minimized while meeting the cooling load demand;

[0175] Set constraints for the optimization objective function;

[0176] The initialization method adopts an improved new population initialization method, which makes the distribution of the population initial solution more uniform and improves the quality of the initial solution.

[0177] A reverse learning strategy is added after the white whale position is updated to avoid the algorithm falling into the local optimum problem;

[0178] Fitness function: The optimization objective function is used as the fitness function;

[0179] Other conditions:

[0180] The population size is 50 individuals;

[0181] The maximum number of iterations is 1000;

[0182] A central air-conditioning chilled water system, comprising an energy consumption modeling module, a cooling load prediction module and an energy-saving optimization module;

[0183] Among them, the energy consumption modeling module establishes the energy consumption model of the chiller and the energy consumption model of the chilled water pump, and the cooling load prediction module establishes the cooling load prediction model; the energy consumption modeling module and the cooling load prediction module transmit data to the energy-saving optimization module, and the energy-saving optimization module processes the received data and outputs the optimization parameters.

[0184] The energy-saving optimization module includes an objective function unit, an optimization algorithm unit and a constraint condition unit. The data of the energy consumption modeling module is transmitted to the objective function unit, and the data of the cooling load prediction module is transmitted to the constraint condition unit. The objective function unit and the constraint condition unit output optimization parameters through the improved Beluga algorithm in the optimization algorithm unit.

[0185] The energy-saving optimization module also includes chiller equipment data, refrigerated water pump equipment data, equipment physical constraints and constraints between equipment. The chiller equipment data and refrigerated water pump equipment data are transmitted to the objective function unit, and the equipment physical constraints and constraints between equipment are transmitted to the constraint condition unit.

[0186] In this embodiment, the energy consumption modeling module collects relevant data on the air-conditioning chiller and the chilled water pump and establishes a relevant equipment energy consumption model, and uses the least square method for parameter identification; the cold load prediction module uses data such as cold load and outdoor environment; the energy-saving optimization module uses the improved Beluga algorithm to optimize equipment operating parameters while meeting the cold load demand, and makes the energy consumption of central air-conditioning chilled water tend to the minimum value. The relevant data collection is to collect the historical energy consumption operation data and equipment operation parameter data of the central air-conditioning chilled water; the historical energy consumption operation data and equipment operation parameter data are quantified and used as parameter solution data for the energy consumption model, and an energy consumption mathematical model is established for the air-conditioning chiller and the chilled water pump, and the least square method is used for parameter identification.

[0187] The data collected by the cooling load prediction module include cooling load data and outdoor meteorological data; wherein, the cooling load data comes from the energy meter installed on the side of the central air-conditioning chiller, and the outdoor meteorological data comes from the data of the outdoor small meteorological station; the cooling load data and the outdoor meteorological data are quantified, and the LSSVM regression algorithm is adopted, and the particle swarm algorithm is adopted to optimize the penalty parameters and kernel parameters to obtain the final cooling load prediction model, and the next point cooling load value is predicted in combination with relevant data.

[0188] Example 2

[0189] Taking the actual data of a certain air-conditioning site as an example, the present invention collects the cooling load data of the energy meter, outdoor temperature data, outdoor humidity data, outdoor irradiance data, chiller energy consumption data, chiller outlet water temperature data, refrigeration pump energy consumption data, chilled water flow, refrigeration pump frequency and cooling water return temperature data as the data source of the entire case, and the collection cycle is 5 minutes. The case analysis is as follows:

[0190] (1) According to step S1, the instantaneous energy consumption data of the refrigerator and the refrigeration pump, the cooling water return temperature, the chilled water supply temperature, the chilled water flow rate and other data are used to construct the energy consumption mathematical model of the refrigerator and the refrigeration pump using the least squares method. The model parameters finally solved are shown in the following formula:

[0191] Mathematical model of freezer:

[0192] 19.938+1.5498*(T co -T wo )-1.3171*(T co -T wo ) 2 -1.2519×10 -2 *Q+0.1339*Q 2

[0193] +3.6465×10 -5 *(T co -T wo )Q

[0194] Mathematical model of refrigeration pump

[0195] 1.4024×10 -2 -0.6884*L+2.6315×10 -3 *L 2 -2.7495×10 -6 *L 3

[0196] (2) According to step S2, a cooling load prediction model PSO-LSSVM is constructed by combining the cooling load data, outdoor temperature data, outdoor humidity data, and outdoor irradiance data of the energy meter. The model input is the cooling load and the outdoor temperature, humidity, and irradiance data. The model output is the cooling load of the next point. The gray correlation analysis method is used to analyze the influencing factors, and the influencing factors with a correlation greater than 0.8 are selected as the input of the model. Finally, the PSO-LSSVM model is trained.

[0197] The parameter settings of PSO are: learning c1=c2=1.5, population size is 50, maximum number of iterations is 1000, fitness function is the mean square error of training data, and the optimization interval of penalty parameter and kernel parameter in LSSVM is [0.01,100]. Since the cold load prediction is the cold load of the next point, in order to verify the superiority of PSO-LSSVM, the present invention saves each prediction in the database, and saves the data of 24 points in total for comparison with the actual data. The specific effect is as follows: Figure 4 As shown. Figure 4 It can be seen that the predicted curve is basically consistent with the actual curve, meeting the basic needs of air conditioning energy saving.

[0198] (3) According to step S3, the energy consumption model obtained in step (1) is combined to construct the objective function of the improved Beluga algorithm:

[0199] minP=min(P ch +P chiled pump )

[0200] =min(19.938+1.5498*(T co -T wo )-1.3171*(T co -T wo ) 2 -1.2519×10 -2 *Q+0.1339*Q 2 +

[0201] 3.6465×10 -5 *(T co -T wo )Q+1.4024×10 -2 -0.6884*L+2.6315×10 -3 *L 2 -2.7495×10 -6 *L 3 )

[0202] The optimization variables and constraints are as follows:

[0203] The cooling tower outlet water temperature selects the maximum and minimum values ​​in actual operation, with the upper and lower limits being 27°C and 23°C respectively.

[0204] The upper and lower limits of the chilled water supply temperature are 12°C and 7°C respectively.

[0205] The upper and lower limits of chilled water flow are 400 m 3 / h and 200 m 3 / h .

[0206] The upper and lower limits of the chilled water supply and return temperature difference are set at 5℃ and 3℃.

[0207] Combined with the cooling load prediction results and the first law of thermodynamics, the following formula requirements are met:

[0208] c·G·Δt=Q

[0209] Among them, c is the specific heat capacity of water; G is the chilled water flow rate; Δt is the supply and return water temperature difference, and Q is the air conditioning cooling load.

[0210] Other conditions:

[0211] For specific algorithm internal parameters, see the above explanation of the basic White Whale algorithm;

[0212] The beluga population size is 50 individuals;

[0213] The maximum number of iterations is 1000;

[0214] In order to verify the energy-saving performance of the improved algorithm, the present invention adds the unimproved BWO algorithm for analysis. The specific comparison of BWO and IBWO before and after energy-saving optimization is shown in the figure below. Figure 5 As shown, Figure 6 is the hourly energy saving rate of the IBWO algorithm.

[0215] according to Figure 5 As can be seen from the figure, the red color represents the original power consumption, the blue + curve is the power consumption diagram of the air conditioning system chilled water optimization using the BWO algorithm, and the black curve is the power consumption after IBWO optimization. Figure 5 It can be seen that the power consumption of the optimization effect diagram of the IBWO algorithm is basically below the power consumption of the original state and the power consumption of the BWO optimization for 24 hours a day, which shows the superiority of the IBWO algorithm in energy saving in the air-conditioning chilled water system.

[0216] Depend on Figure 6 It can be seen that the energy saving rate is higher in the morning and evening, with the highest being close to 18%, while the energy saving rate is lower at noon, which is also in line with the actual situation on site. The temperature at noon is higher and the demand for cooling load is high, so the energy saving rate is relatively low. As the temperature drops in the morning and evening, the demand for cooling load decreases, and the relative energy saving rate increases. In summary, the energy saving rate is between 10% and 18%, and the average energy saving rate is around 13.74%, which shows the effectiveness of the IBWO algorithm in air conditioning energy saving.

[0217] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. An energy-saving optimization control method for a central air-conditioning chilled water system, characterized in that: The energy saving optimization method The process includes: S1. Establish energy consumption model of central air-conditioning chilled water system; S2. Firstly, the grey correlation analysis method is used to analyze the correlation of the selected cooling load influencing factors, and the strongly correlated influencing factors are screened out, and a central air-conditioning cooling load prediction model based on PSO-LSSVM is established; S3. Use the improved Beluga algorithm to meet the prediction of the next point cooling load value, optimize the energy consumption model parameters of the central air-conditioning chilled water system, and make the output result of the energy consumption model reach the minimum value; The energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model periodically update data, predict the cooling load through the real-time updated energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model, and optimize the energy consumption model at the same time.

2. The energy-saving optimization control method for a central air-conditioning chilled water system according to claim 1, characterized in that: The establishment of the energy consumption model includes the following processes: S11, collecting historical energy consumption data and equipment operation parameter data of the central air-conditioning chilled water system as data for energy consumption model parameter identification; S12. Establish a mathematical model of equipment energy consumption of the central air-conditioning chilled water system, and use the least squares method to identify parameters based on the collected historical energy consumption data and equipment operation parameter data.

3. The energy-saving optimization control method for a central air-conditioning chilled water system according to claim 2, characterized in that: The historical energy consumption operation data includes the instantaneous power of the chiller and the instantaneous power of the chilled water pump; the equipment operation parameter data includes the cooling load of the chiller, the chilled water supply temperature, the chilled water return temperature, the cooling water return temperature, the chilled water pump operation frequency and the total chilled water flow.

4. The energy-saving optimization control method for a central air-conditioning chilled water system according to claim 1, characterized in that: The establishment of the central air conditioning cooling load prediction model includes the following processes: S21, collecting central air-conditioning cooling load data and outdoor environment data as historical data for cooling load prediction model training; S22, using the collected historical data as training samples for the cooling load prediction model, and using the particle swarm algorithm to perform penalty parameter and kernel function optimization training to obtain the final cooling load prediction model as the terminal constraint for the optimization of the central air-conditioning chilled water system; S23, determining the prediction cycle, using the current cooling load data and the current outdoor environment data actually taken, as well as the cooling load data and the outdoor environment data of the previous point as inputs of the cooling load prediction model, and predicting the cooling load value of the next point.

5. The energy-saving optimization control method for a central air-conditioning chilled water system according to claim 4, characterized in that: The outdoor environment data includes outdoor meteorological data, and the outdoor meteorological data includes humidity and irradiance parameters.

6. The energy-saving optimization control method for a central air-conditioning chilled water system according to claim 1, characterized in that: The energy consumption model of the central air-conditioning chilled water system and the central air-conditioning cooling load prediction model update data include the following process: S31, determine the optimization parameters of the central air-conditioning chilled water system, collect the real-time operation energy consumption data, equipment operation parameter data and outdoor environment data of the central air-conditioning chilled water system, and use the improved Beluga algorithm to solve the optimal system operation parameters while satisfying the predicted cooling load value of the next point; S32, outputting the equipment operation parameter data of the final central air-conditioning chilled water system; S33, determining whether the optimization cycle time of the central air-conditioning chilled water system has been reached, and repeating S31 to S33 until the optimization is stopped; S34, determining the update time of the energy consumption model and the cooling load prediction model, and if the update time is reached, updating the energy consumption model parameters and the cooling load prediction model and storing them; Among them, if the optimal cycle time of the central air-conditioning chilled water system is reached, the next point cooling load value, parameter optimization and output optimization structure are predicted.

7. The energy-saving optimization control method for a central air-conditioning chilled water system according to claim 1, characterized in that: The improvement of the White Whale algorithm includes the following process: S101, initializing IBWO parameters, including the number of iterations and population size; S102, initializing the population using a new population initialization strategy; S103, calculating the fitness values ​​of individual beluga whales and selecting the individual with the best fitness; S104. Calculate the balance factor B j and whale fall probability W j ; S105, if B j If it is less than 0.5, it is in the development stage, otherwise it is in the exploration stage; S106, if B j Greater than W j , it is the whale fall stage, otherwise it is the reverse learning strategy; S107, updating fitness to determine the best individual; S108. If the maximum number of iterations is reached, the parameter adjustment result is output, otherwise S103 is repeated until the maximum number of iterations is reached.

8. A central air conditioning chilled water system using any one of claims 1 to 7, characterized in that: It includes energy consumption modeling module, cooling load prediction module and energy saving optimization module; Among them, the energy consumption modeling module establishes the energy consumption model of the chiller and the energy consumption model of the chilled water pump, and the cooling load prediction module establishes the cooling load prediction model; the energy consumption modeling module and the cooling load prediction module transmit data to the energy-saving optimization module, and the energy-saving optimization module processes the received data and outputs the optimization parameters.

9. A central air-conditioning chilled water system according to claim 8, characterized in that: The energy-saving optimization module includes an objective function unit, an optimization algorithm unit and a constraint condition unit. The data of the energy consumption modeling module is transmitted to the objective function unit, and the data of the cooling load prediction module is transmitted to the constraint condition unit. The objective function unit and the constraint condition unit output optimization parameters through the improved Beluga algorithm in the optimization algorithm unit.

10. A central air-conditioning chilled water system according to claim 9, characterized in that: The energy-saving optimization module also includes chiller equipment data, refrigerated water pump equipment data, equipment physical constraints and constraints between equipment. The chiller equipment data and refrigerated water pump equipment data are transmitted to the objective function unit, and the equipment physical constraints and constraints between equipment are transmitted to the constraint condition unit.

Images