Energy-saving control method and system based on dynamic air conditioning operation data, and medium

Abstract

This invention provides an energy-saving control method, system, and medium based on dynamic air conditioning operation data. The method includes: collecting historical air conditioning operation data as training samples; establishing an air conditioning load prediction model and an air conditioning system simulation model using a neural network algorithm; and then, using minimum system energy consumption as the optimization objective, employing a branch-and-bound method to dynamically optimize the system's operating state while ensuring safety. This method solves the problem of traditional central air conditioning control systems, which fragment the entire system and use PID or single-factor adjustment for independent control of only a single unit or group of equipment, resulting in the inability to find the global optimum. This approach maximizes the energy-saving potential of the air conditioning system.

Description

Technical Field

[0001] This invention relates to the technical field of air conditioning control algorithms, specifically to an energy-saving control method, system, and medium based on dynamic air conditioning operation data. Background Technology

[0002] Currently, most air conditioning systems in medical and health buildings are still operated manually. The few existing group control systems employ traditional PID control, but due to the strong coupling, large delays, and high inertia of these systems, the control performance is unsatisfactory. Dynamic optimization of chiller group control is a typical mixed-integer programming (MINLP) problem, involving not only the optimization of integer variables such as equipment start-up and shutdown, but also the optimization of continuous variables such as equipment operating parameters. These characteristics make it difficult to establish accurate mathematical models for such systems. Therefore, establishing a dynamic optimization control strategy for air conditioning systems is crucial.

[0003] Currently, intelligent algorithms for dynamic control of air conditioning systems utilize physical models to optimize global total energy consumption as the objective function. Each iteration eliminates "bad solutions" and generates "new solutions," then solves a high-dimensional optimization model to obtain effective optimized control parameters. While the "optimal solution" obtained by this method has low total energy consumption, it fails to fully consider the comfort requirements of the optimized end-user. For example, Chinese patent document CN114543273A discloses an adaptive deep learning optimization energy-saving control algorithm for centralized air conditioning cooling systems. However, the optimal solution obtained by this algorithm does not fully consider the comfort requirements of the optimized end-user. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this invention is to provide an energy-saving control method, system, and medium based on dynamic air conditioning operation data.

[0005] An energy-saving control method based on dynamic air conditioning operation data provided by the present invention includes:

[0006] Step S1: Collect historical air conditioning operation data as training samples, and use neural network algorithms to establish an air conditioning load prediction model and an air conditioning system simulation model;

[0007] Step S2: Record the control parameters of the current status of each device in the air conditioning system, and generate an initial device status matrix ST based on the control parameters. n×1 ;

[0008] Step S3: With ST n×1 The output parameters of each device status shown in the matrix are used as the starting position for the search. The branch and bound method is used to generate the workable solution matrix BD. n×m ;

[0009] Step S4: BDn×m Assign values ​​to the device state matrix ST row by row. n×1 The air conditioning system simulation model is used to determine whether the operating scheme corresponding to the current equipment state matrix is ​​dangerous. If so, BD is selected. n×m Assign values ​​to the next set of data in the matrix; if it does not exist, proceed to step S5.

[0010] Step S5: Calculate the energy consumption F of the current operating scheme using the air conditioning system simulation model. ec The indoor comfort level F of the current operating scheme is calculated using an air conditioning load prediction model. co Energy consumption F ec and indoor comfort F co The indicators are weighted and the optimal solution ST′ is selected from all feasible solutions. best Issued to on-site equipment for execution.

[0011] Preferably, it further includes:

[0012] Step S6: When executing the optimal solution ST′ best After the set time has elapsed, proceed to step S1 for the next round of optimization.

[0013] Preferably, the method for establishing the air conditioning load prediction model includes:

[0014] Step S1.1: Collect samples, filter, and normalize the historical data of on-site equipment operation and outdoor environmental data;

[0015] Step S1.2: Establish an air conditioning load prediction model f using a neural network algorithm. The air conditioning load prediction model is used to describe the statistical relationship between various environmental influencing factors and the change in indoor average temperature.

[0016] Step S1.3: Solve for the inverse function f of the air conditioning load prediction model f. -1 According to the inverse function f -1 And the cooling capacity required to solve for known environmental impact factors.

[0017] Preferably, the environmental impact factors include outdoor temperature and humidity, time period, supply and return water temperature and flow rate, and cooling capacity.

[0018] Preferably, the air conditioning system simulation model is established based on a neural network algorithm to establish a statistical model of environmental impact factors and target quantities. The air conditioning system simulation model includes a return water temperature model, a room temperature change rate model, a refrigeration unit cooling capacity, energy consumption, and efficiency model, a water pump flow rate and energy consumption model, and a cooling tower outlet water speed and energy consumption model.

[0019] Preferably, the energy consumption F of the current operating scheme ec The calculation methods include:

[0020] F ec =∑F wcu +∑F ldp +∑F lqp +∑F ct

[0021] In the formula, F wcu For the energy consumption of each cooling unit, F ldp For the energy consumption of each chilled water pump, F lqp For the energy consumption of each cooling water pump, F ct Energy consumption of each cooling tower;

[0022] F wcu =f -1 (T sw ,d sw ,T hs ,q hs ,T gs-set ,q lq ,T lq )

[0023] In the formula, T sw outdoor temperature, d sw For outdoor humidity, T hs q represents the return water temperature. hs For return water flow rate, T gs-set q is the setpoint for the water supply temperature. lq T represents the cooling water flow rate. lq This refers to the cooling water temperature.

[0024] F ldp =f -1 (f ld-pl ,I fm-sta ,f other-ldpl )

[0025] In the formula, f ld-pl I is the frequency of the chilled water pump. fm-sta For valve status, f other-ldpl For other chilled water pump status;

[0026] F lqp =f -1 (f lq-pl ,I fm-sta ,f other-lqpl )

[0027] In the formula, f lq-pl For the cooling water pump frequency, I fm-sta For valve status, f other-lqpl For other cooling water pump status;

[0028] F ct =f-1 (T sw ,d sw ,T hs ,q hs ,I fj-sta )

[0029] In the formula, T sw outdoor temperature, d sw For outdoor humidity, T hs q represents the return water temperature. hs For return water flow rate, I fj-set Set the fan status.

[0030] Preferably, the indoor comfort level F of the current solution is... co The calculation methods include:

[0031] Step S5.1: Calculate the required cooling capacity Q within the preset time using the air conditioning load prediction model. yc ;

[0032] Step S5.2: Using the refrigeration capacity model in the air conditioning system simulation model, simulate and calculate the equipment according to the state matrix ST. n×1 The total cooling capacity Q of the chiller after running for a preset time mn ;

[0033] Step S5.3: Calculate the predicted cooling capacity Q based on the load. yc The total cooling capacity Q simulated by the refrigeration mechanism cooling capacity model mn By comparison, the indoor comfort index F was calculated. co :

[0034]

[0035] An energy-saving control system based on dynamic air conditioning operation data provided by the present invention includes:

[0036] Module M1: Collects historical air conditioning operation data as training samples, and uses neural network algorithms to establish an air conditioning load prediction model and an air conditioning system simulation model;

[0037] Module M2: Records the control parameters of the current status of each device in the air conditioning system, and generates the initial device state matrix ST based on the control parameters. n×1 ;

[0038] Module M3: with ST n×1 The output parameters of each device status shown in the matrix are used as the starting position for the search. The branch and bound method is used to generate the workable solution matrix BD. n×m ;

[0039] Module M4: (This likely refers to a module or module) n×m Assign values ​​to the device state matrix ST row by row.n×1 The air conditioning system simulation model is used to determine whether the operating scheme corresponding to the current equipment state matrix is ​​dangerous. If so, BD is selected. n×m The next set of data in the matrix is ​​assigned a value; if it does not exist, proceed to module M5.

[0040] Module M5: Calculates the energy consumption F of the current operating scheme using an air conditioning system simulation model. ec The indoor comfort level F of the current operating scheme is calculated using an air conditioning load prediction model. co Energy consumption F ec and indoor comfort F co The indicators are weighted and the optimal solution ST′ is selected from all feasible solutions. best Issued to on-site equipment for execution.

[0041] Preferably, it further includes:

[0042] Module M6: When executing the optimal solution ST′ best After a set time, the process proceeds to module M1 for the next round of optimization. According to the present invention, a computer-readable storage medium storing a computer program, when executed by a processor, implements the steps of an energy-saving control method based on dynamic air conditioning operating data.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. This algorithm uses the branch and bound method to perform global optimization of each controllable device in the air conditioning system. This solves the problem that traditional central air conditioning control systems cannot find the global optimum because they only use PID or single-factor adjustment to independently control a single device or a single group of devices, which is a fragmented approach to the whole system. This algorithm maximizes the energy-saving potential of the air conditioning system.

[0045] 2. By directly using the real-time collected indoor temperature and humidity parameters to control the air conditioning system, the problems of lag and poor control effect of traditional PID control are solved. While improving the comfort of the air conditioning terminal, it also reduces the energy consumption of the air conditioning and avoids overcooling or overheating.

[0046] 3. By continuously collecting new operating data samples of the air conditioning system to replace the old samples, and using spare time to train the neural network model, the problem that the model would gradually deviate from the actual situation over time was solved, giving the model a strong adaptive ability and enhancing the robustness of the algorithm.

[0047] Other beneficial effects of the present invention will be explained in detail through the introduction of specific technical features and technical solutions in specific embodiments. Those skilled in the art should be able to understand the beneficial technical effects brought about by these technical features and technical solutions through the introduction of these technical features and technical solutions. Attached Figure Description

[0048] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0049] Figure 1 This is a flowchart of the energy-saving control method based on dynamic air conditioning operation data in this invention;

[0050] Figure 2 This is a diagram of the air conditioning system prediction model in this invention;

[0051] Figure 3 This is a schematic diagram showing the change of PMV at 25°C over time in this invention;

[0052] Figure 4 This is a schematic diagram showing the change of PMV at 26°C over time in this invention;

[0053] Figure 5 This is a schematic diagram showing the change of PMV at 27°C over time in this invention;

[0054] Figure 6 This is a schematic diagram showing the change of PMV at 28°C over time in this invention;

[0055] Figure 7 This is a schematic diagram showing the change of PMV at 29°C over time in this invention;

[0056] Figure 8 This is a simulated outdoor temperature map for a given day in an embodiment of the present invention;

[0057] Figure 9 This is a schematic diagram showing the number of iterations and the effect of the energy-saving algorithm model of the present invention. Detailed Implementation

[0058] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0059] This invention discloses an energy-saving control method based on dynamic air conditioning operation data. The method first utilizes historical operating data of the air conditioning system and establishes an air conditioning load prediction model and an air conditioning system simulation model based on a neural network algorithm. Then, with the minimum system energy consumption as the optimization objective, the branch-and-bound method is used to dynamically optimize the system operating state while ensuring safety.

[0060] Specifically, refer to Figure 1As shown, the energy-saving control method based on dynamic air conditioning operation data includes:

[0061] Step S1: Collect historical air conditioning operation data as training samples, and use neural network algorithms to establish an air conditioning load prediction model and an air conditioning system simulation model.

[0062] Step S2: Record the control parameters of the current status of each device in the air conditioning system, and generate an initial device status matrix ST based on the control parameters. n×1 The control parameters include the on / off status of each chiller, the set temperature of the water supply, the on / off status of each water pump, the water pump frequency, the on / off status of each cooling tower, and the cooling tower fan frequency.

[0063] Step S3: With ST n×1 The output parameters of each device status shown in the matrix are used as the starting position for the search. The branch and bound method is used to generate the workable solution matrix BD. n×m ;

[0064] Step S4: BD n×m Assign values ​​to the device state matrix ST row by row. n×1 The air conditioning system simulation model is used to determine whether the current operating scheme corresponding to the equipment state matrix is ​​dangerous (e.g., causing the chiller flow rate to be lower than the manufacturer's set value). If so, BD is selected. n×m Assign values ​​to the next set of data in the matrix; if it does not exist, proceed to step S5.

[0065] Step S5: Calculate the energy consumption F of the current operating scheme using the air conditioning system simulation model. ec The indoor comfort level F of the current operating scheme is calculated using an air conditioning load prediction model. co Energy consumption F ec and indoor comfort F co The indicators are weighted and the optimal solution ST′ is selected from all feasible solutions. best Issued to on-site equipment for execution.

[0066] In a preferred embodiment, step S6 is further included: when the optimal solution ST′ is executed. best After the set time has elapsed, proceed to step S1 for the next round of optimization.

[0067] The method for establishing the air conditioning load prediction model includes:

[0068] Step S1.1: Perform sample collection, filtering, and normalization operations on the historical data of on-site equipment operation and outdoor environmental data;

[0069] Step S1.2: Establish an air conditioning load prediction model f using a neural network algorithm. This model describes the statistical relationship between various environmental influencing factors and changes in indoor average temperature. The environmental influencing factors include outdoor temperature and humidity, time period, supply and return water temperature and flow rate, and cooling capacity. (Refer to...) Figure 2 The above refers to the outdoor temperature and humidity, x2 is the time period, x3 is the cooling capacity, x4 is the supply and return water temperature and flow rate, and y is the change in the indoor average temperature.

[0070] Step S1.3: Solve for the inverse function f of the air conditioning load prediction model f. -1 According to the inverse function f -1 Given the outdoor temperature and humidity, time period, and required change in indoor average temperature, the required cooling capacity for a future time window can be calculated.

[0071] The air conditioning system simulation model is established based on a neural network algorithm. A statistical model of environmental impact factors and target quantities is built, following steps S1.1-S1.3. The air conditioning system simulation model includes a return water temperature model, a room temperature change rate model, a refrigeration unit cooling capacity, energy consumption, and efficiency model, a water pump flow rate and energy consumption model, and a cooling tower outlet water velocity and energy consumption model. The neural network input / output layers for each model are shown in the table below:

[0072] Table 1

[0073]

[0074] Energy consumption F of the current operating scheme ec The detailed calculation process is as follows:

[0075] F ec =∑F wcu +∑F ldp +∑F lqp +∑F ct

[0076] In the above formula, F wcu Energy consumption of each cooling unit; F ldp Energy consumption of each chilled water pump; F lqp Energy consumption of each cooling water pump; F ct Energy consumption of each cooling tower;

[0077] Among them, F wcu =f -1 (T sw ,d sw ,T hs ,q hs ,T gs-set ,q lq ,T lq )

[0078] In the formula, T sw outdoor temperature; d sw Outdoor humidity; T hs q represents the return water temperature. hs T represents the return water flow rate. gs-set The setpoint for the water supply temperature; q lq T is the cooling water flow rate; lq This refers to the cooling water temperature.

[0079] A neural network algorithm is used to establish a chiller energy consumption model f (the model input layers are: outdoor temperature and humidity, return water temperature, return water flow rate, supply water temperature setting, cooling water flow rate, and cooling water temperature; the output layer is: chiller energy consumption). The energy consumption of each chiller can be obtained by inverting this model.

[0080] Similarly, neural network algorithms were used to construct energy consumption models for chilled water pumps, cooling water pumps, and cooling towers, respectively. The input and output layers are shown in Table 1. Inverting these models yields:

[0081] F ldp =f -1 (f ld-pl ,I fm-sta ,f other-ldpl )

[0082] In the formula, f ld-pl Frequency of chilled water pump; I fm-sta Valve status; f other-ldpl This indicates the status of other chilled water pumps.

[0083] F lqp =f -1 (f lq-pl ,I fm-sta ,f other-lqpl )

[0084] In the formula, f lq-pl For cooling water pump frequency; I fm-sta Valve status; f other-lqpl This indicates the status of other cooling water pumps.

[0085] F ct =f -1 (T sw ,d sw ,T hs ,q hs ,I fj-sta )

[0086] In the formula, T sw outdoor temperature; d sw Outdoor humidity; T hs q represents the return water temperature. hs For return water flow rate; Ifj-set Set the fan status.

[0087] The current plan's indoor comfort level F co The detailed calculation process is as follows:

[0088] Step 5.1: Calculate the required cooling capacity Q over a certain period of time using the air conditioning load prediction model. yc ;

[0089] Step 5.2: Using the refrigeration capacity model of the air conditioning system simulation model (model input and output layers are shown in Table-1), simulate and calculate the equipment according to the state matrix ST. n×1 After running for a period of time, the total cooling capacity Q of the chiller mn ;

[0090] Step 5.3: Calculate the predicted cooling capacity Q. yc The total cooling capacity Q simulated by the refrigeration mechanism cooling capacity model mn Compare the results and calculate the indoor comfort index F using the following formula. co :

[0091]

[0092] The air conditioning energy-saving algorithm in this invention will be further explained below.

[0093] 1. Establish a power consumption model

[0094] Establish a functional relationship between air conditioning energy consumption and changes in indoor and outdoor temperatures. The energy consumption of the first unit is related to the current indoor temperature, the indoor temperature at the last time point, and the current outdoor ambient temperature as follows.

[0095]

[0096] Assuming the energy consumed by the air conditioners in the first hour is a fixed value, the total energy consumed by all air conditioners in the second hour is as follows:

[0097]

[0098]

[0099] The total energy consumption function of all air conditioners throughout the day is set as a quadratic function (3), where c is a constant whose value is related to electricity price and energy consumption. For the i-th air conditioner, the total energy consumed in a day is as follows:

[0100] C i =c[E1E2…E n ][P i,1 P i,2 …P i,n ]′ (4)

[0101] Ci =c(E1P i,1 +E2P i,2 +…+E n P i,n (5)

[0102] The total energy consumed by an air conditioner in a day will increase or decrease depending on the set temperature. According to formula (5), the total energy consumed by an air conditioner in a day is only related to its own energy consumption and that of other air conditioners. Therefore, its function can also be expressed as:

[0103]

[0104] The transformation yields:

[0105] According to equations (1) and (4), the total energy consumed by the i-th air conditioner in a day can be expressed as the derivative of the current temperature value:

[0106]

[0107] Combining equations (6) and (7), we can obtain:

[0108]

[0109] Based on formula (1), we can obtain:

[0110]

[0111]

[0112] Combining equations (8), (9), and (10) and replacing them with (7), we can obtain:

[0113]

[0114] According to (12), the total energy consumed by the i-th air conditioner in a day will change with the temperature set by the air conditioner.

[0115] 2. Select thermal comfort index

[0116] Thermal comfort control is a method of controlling air conditioning based on the thermal comfort index felt by the human body in a thermal environment. PMV, as a subjective evaluation standard for human thermal sensation, categorizes human thermal sensation into seven levels based on different degrees, using the thermal comfort index as the indoor comfort level (F). co Evaluation indicators. Based on the impact of air conditioner set temperature on PMV values, it is clear that temperature has a significant impact on thermal comfort. For example, referring to... Figure 3-7As shown, for indoor set temperatures ranging from 25°C to 29°C, when the indoor temperature is set to 25°C, the PMV value ranges from -0.2 to 0.5 within a day. When the indoor temperature is set to 29°C, the PMV value ranges from 1.0 to 2.5 within a day. On the other hand, even at the same set temperature and at different times, indoor thermal comfort fluctuates significantly.

[0117] 3. Optimization of Air Conditioning Energy-Saving Model Based on Genetic Algorithm

[0118] Initialize the overall

[0119] Within the aforementioned temperature range, a certain number of chromosomes are randomly generated, and chromosomes that do not meet the conditions are eliminated. The initial population size should be determined based on the actual situation of the problem. Since the integer solution range of the air conditioning power consumption model is the set temperature range, i.e., 5, and only 12 hours are controlled per day, the required range size is 5^12. Therefore, an initial population size of less than 100 can satisfy the optimization conditions and is highly efficient.

[0120] Fitness function design

[0121] Since the objective function is the solution to a minimum problem, the function value range [10.036, 17.751] is a fixed interval of positive numbers, meaning the energy consumption values ​​are obtained from the highest and lowest temperatures in the air conditioning solution set. Furthermore, indoor comfort conditions must also be considered, i.e., the PMV value obtained through testing. Since the PMV value ranges from [-3, 3], to ensure the most comfortable indoor environment possible, the indoor PMV value must be within [-1, 1].

[0122] 4. MATLAB simulation of the optimization model

[0123] In the simulation experiment, meteorological data of the region in July during the summer were selected as parameters, including the region's summer temperature, humidity, wind speed, average thermal radiation temperature, human activities, clothing, etc.

[21] . The outdoor temperature on the simulated day was as follows: Figure 8 As shown, the relative humidity is 93%. The wind speed is very gentle. The ultraviolet radiation is very strong. The human energy metabolism setting is 58W / m. 2 The thermal resistance of the clothing is 0.5clo. The simulation settings are as follows: The number of air conditioners in the system is m = 200 (the following analysis considers only one air conditioner; the set temperature values ​​for the other air conditioners are randomly generated). The operating time of the air conditioners is 12 hours per day, corresponding to the hourly values ​​from 7:00 to 19:00. The temperature range is set to integer values ​​within the range of [25℃, 29℃]. In the air conditioner power consumption model, the constants are α = 0.9, β = -10, and c = 0.6.

[0124] Since the ranges of parameters m and n are defined as 0° ≤ °m ≤ 17 and 0° < °n ≤ 17, a certain relationship exists between the two values. When the value of m increases, the impact of energy-saving effect on the algorithm decreases, while the impact of the thermal comfort index on the algorithm increases. The same result occurs when n increases.

[0125] First, the initial population size significantly affects the number of generations to convergence. A larger total number of generations results in faster convergence. However, within a suitable range, changes in the crossover probability Pc, mutation probability Pa, and maximum number of generations have little impact on the algorithm's optimization performance. Therefore, in this study, the genetic algorithm's parameters were set as follows: population size 100, crossover probability Pc 0.5, mutation probability Pm 0.01, and maximum number of generations 100.

[0126] Secondly, when the constant parameters m and n in the fitness function formula are set to m=1, n=1, m=4, n=2, or m=6, n=3, although good energy efficiency can be achieved, the indoor thermal comfort index is poor and cannot meet the user's comfort requirements. Therefore, these three pairs of values ​​are not suitable. In contrast, when the constant parameters m and n are set to m=8, n=4 and m=10, n=4, although the energy-saving effect is relatively poor, a comfortable indoor environment can be provided for the user. Therefore, these two pairs of values ​​meet the design goals of the air conditioning energy-saving method in the study and can be adopted. In this study, the parameters m and n were set to m=8 and n=4.

[0127] Secondly, after determining the constant parameter values ​​of the energy-saving algorithm, the energy-saving effect of the proposed algorithm was analyzed in depth based on the set parameter values. The indoor temperature setpoint optimized by the genetic algorithm at the corresponding time point corresponds to a total daily energy consumption (in yuan) C1 = 12.756. Based on the average value of the human body's suitable temperature range, setting the indoor temperature to 27℃ is equivalent to a total daily energy consumption (in yuan) C1 = 13.675. The randomly generated indoor temperature setpoint corresponds to a total daily energy consumption (in yuan) C1 = 14.286. The energy consumption of the optimized set temperature can be reduced by about 89% compared with the randomly set temperature, and by about 93% compared with the average human body's set temperature range, indicating that the relative thermal comfort of the indoor environment is more ideal. Figure 9 This is a curve showing the change in energy consumption for each generation in the genetic algorithm simulation. As can be seen from the graph, the model converges around the 53rd generation, demonstrating a significant optimization effect.

[0128] This invention also discloses an energy-saving control system based on dynamic air conditioning operation data, comprising:

[0129] Module M1: Collects historical air conditioning operation data as training samples, and uses neural network algorithms to establish an air conditioning load prediction model and an air conditioning system simulation model;

[0130] Module M2: Records the control parameters of the current status of each device in the air conditioning system, and generates the initial device state matrix ST based on the control parameters. n×1 ;

[0131] Module M3: with ST n×1 The output parameters of each device status shown in the matrix are used as the starting position for the search. The branch and bound method is used to generate the workable solution matrix BD. n×m ;

[0132] Module M4: (This likely refers to a module or module) n×m Assign values ​​to the device state matrix ST row by row. n×1 The air conditioning system simulation model is used to determine whether the operating scheme corresponding to the current equipment state matrix is ​​dangerous. If so, BD is selected. n×m The next set of data in the matrix is ​​assigned a value; if it does not exist, proceed to module M5.

[0133] Module M5: Calculates the energy consumption F of the current operating scheme using an air conditioning system simulation model. ec The indoor comfort level F of the current operating scheme is calculated using an air conditioning load prediction model. co Energy consumption F ec and indoor comfort F co The indicators are weighted and the optimal solution ST′ is selected from all feasible solutions. best Issued to on-site equipment for execution.

[0134] Module M6: When executing the optimal solution ST′ best After a set time, the system will proceed to module M1 for the next round of optimization.

[0135] In a preferred embodiment, the method for establishing the air conditioning load prediction model includes:

[0136] Module M1.1: Performs sample collection, filtering, and normalization processing on historical data of field equipment operation and outdoor environmental data;

[0137] Module M1.2: An air conditioning load prediction model f is established using a neural network algorithm. The air conditioning load prediction model is used to describe the statistical relationship between various environmental influencing factors and the change in indoor average temperature. The environmental influencing factors include outdoor temperature and humidity, time period, supply and return water temperature and flow rate, and cooling capacity.

[0138] Module M1.3: Solving for the inverse function f of the air conditioning load forecasting model f -1 According to the inverse function f -1 And the cooling capacity required to solve for known environmental impact factors.

[0139] The present invention also discloses a computer-readable storage medium storing a computer program, which may be a portable hard drive, a USB flash drive, etc. When the computer program is executed by a processor, it implements the steps of an energy-saving control method based on dynamic air conditioning operation data.

[0140] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0141] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for energy saving control based on dynamic air conditioning operation data, characterized by, include: Step S1: Collect historical air conditioning operation data as training samples, and use neural network algorithms to establish an air conditioning load prediction model and an air conditioning system simulation model; Step S2: record the control parameters of the current state of each device of the air conditioning system, and generate an initial device state matrix ST based on the control parameters n×1 ; Step S3: ST n×1 The branch and bound method is used to generate the executable scheme matrix BD with the device state output parameters of each device shown in the matrix as the search starting position n×m ; Step S4: assign the BD n×m matrix to the device state matrix ST n×1 row by row, and determine whether the operation scheme corresponding to the current device state matrix exists danger by using the air conditioning system simulation model. If it exists, select the next group of data in the BD n×m matrix to assign; if it does not exist, go to step S5. Step S5: Calculate the energy consumption F of the current operation scheme using the air conditioning system simulation model ec , Calculate the indoor comfort F of the current operation scheme using the air conditioning load prediction model co , Calculate the energy consumption F ec and the indoor comfort F co index, and select the optimal scheme ST best from all feasible schemes by weighted calculation Downlink to on-site equipment for execution; the indoor comfort of the current operating scenario F co The calculation method comprises: Step S5.1: calculating the required refrigerating capacity in the preset time by using the air conditioning load prediction model Q yc ; Step S5.2: simulate the total refrigerating capacity of the cold machine according to the state matrix ST of the equipment by using the cold machine refrigerating capacity model in the air conditioning system simulation model n×1 Total refrigerating capacity of the cold machine after running for a preset time Q mn ; Step S5.3: Predicting the cooling load Q yc Total refrigeration capacity simulated by the cold machine refrigeration capacity model Q mn Comparing, calculating the indoor comfort index F co : ； The input layer of the air conditioning load prediction model includes outdoor temperature and humidity, time period, cooling capacity, and supply and return water temperature and flow rate. The output layer of the air conditioning load prediction model is the change in average indoor temperature. The air conditioning system simulation model includes a return water temperature model, a room temperature change rate model, a refrigeration unit cooling capacity, energy consumption, and efficiency model, a water pump flow rate and energy consumption model, and a cooling tower outlet water velocity and energy consumption model.

2. The energy saving control method based on dynamic air conditioning operation data according to claim 1, characterized in that, Also includes: Step S6: When executing the optimal solution ST best After the set time has elapsed, proceed to step S1 for the next round of optimization.

3. The energy-saving control method based on dynamic air conditioning operation data according to claim 1, characterized in that, The method for establishing the air conditioning load prediction model includes: Step S1.1: Collect samples, filter, and normalize the historical data of on-site equipment operation and outdoor environmental data; Step S1.2: Establish an air conditioning load prediction model using a neural network algorithm. f The air conditioning load prediction model is used to describe the statistical relationship between various environmental influencing factors and the change in indoor average temperature; Step S1.3: Solve the air conditioning load prediction model f inverse function f -1 According to the inverse function f -1 And the cooling capacity required to solve for known environmental impact factors.

4. The energy-saving control method based on dynamic air conditioning operation data according to claim 3, characterized in that, The environmental impact factors include outdoor temperature and humidity, time period, supply and return water temperature and flow rate, and cooling capacity.

5. The energy-saving control method based on dynamic air conditioning operation data according to claim 1, characterized in that, The air conditioning system simulation model is established based on a neural network algorithm, and a statistical model of environmental impact factors and target quantities is established. The air conditioning system simulation model includes a return water temperature model, a room temperature change rate model, a refrigeration unit cooling capacity, energy consumption, and efficiency model, a water pump flow rate and energy consumption model, and a cooling tower outlet water speed and energy consumption model.

6. The energy-saving control method based on dynamic air conditioning operation data according to claim 3, characterized in that, the energy consumption F of the current operating scheme ec The calculation method comprises: In the formula, F wcu For the energy consumption of each cooling unit, F ldp Energy consumption of each chilled water pump F lqp Energy consumption of each cooling water pump F ct Energy consumption of each cooling tower; In the formula, T sw Outdoor temperature d sw Outdoor humidity, T hs The return water temperature, q hs For return water flow rate, T gs-set Set the water supply temperature. q lq For cooling water flow rate, T lq This refers to the cooling water temperature. In the formula, f ld-pl For chilled water pump frequency, I fm-sta Valve status. f other-ldpl For other chilled water pump status; In the formula, f lq-pl For the frequency of the cooling water pump, I fm-sta Valve status. f other-lqpl For other cooling water pump status; In the formula, T sw Outdoor temperature d sw Outdoor humidity, T hs The return water temperature, q hs For return water flow rate, I fj-set Set the fan status.

7. An energy-saving control system based on dynamic air conditioning operation data, characterized in that, include: Module M1: Collects historical air conditioning operation data as training samples, and uses neural network algorithms to establish an air conditioning load prediction model and an air conditioning system simulation model; Module M2: record the control parameters of the current state of each device of the air conditioning system, generate an initial device state matrix ST based on the control parameters n×1 ; Module M3: in ST n×1 The matrix shows each device state output parameter as a search starting position, and a branch and bound method is used to generate a runnable scheme matrix BD n×m ; Module M4: Assigning BD n×m The matrix is assigned to the device state matrix ST row by row n×1 , and the simulation model of the air conditioning system is used to determine whether the operation scheme corresponding to the current device state matrix is dangerous. If it is dangerous, the next set of data in the matrix is selected as BD n×m ; if it is not dangerous, module M5 is entered. Module M5: Calculates the energy consumption F of the current operating scheme using an air conditioning system simulation model. ec The indoor comfort level F of the current operating scheme is calculated using an air conditioning load prediction model. co Energy consumption F ec and indoor comfort F co The indicators are weighted and the optimal solution ST is selected from all feasible solutions. best Issued to on-site equipment for execution; Indoor comfort level of the current operating plan F co The calculation methods include: Step S5.1: Calculate the required cooling capacity within the preset time using the air conditioning load prediction model. Q yc ; Step S5.2: Using the refrigeration capacity model in the air conditioning system simulation model, simulate and calculate the equipment according to the state matrix ST. n×1 Total cooling capacity of the chiller after running for a preset time Q mn ; Step S5.3: Predict the cooling capacity based on the load. Q yc Total cooling capacity simulated by the refrigeration system cooling capacity model Q mn Compare and calculate indoor comfort index F co : ； The input layer of the air conditioning load prediction model includes outdoor temperature and humidity, time period, cooling capacity, and supply and return water temperature and flow rate. The output layer of the air conditioning load prediction model is the change in average indoor temperature. The air conditioning system simulation model includes a return water temperature model, a room temperature change rate model, a refrigeration unit cooling capacity, energy consumption, and efficiency model, a water pump flow rate and energy consumption model, and a cooling tower outlet water velocity and energy consumption model.

8. The energy-saving control system based on dynamic air conditioning operation data according to claim 7, characterized in that, Also includes: Module M6: When executing the optimal solution ST best After a set time, the system will proceed to module M1 for the next round of optimization.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the energy-saving control method based on dynamic air conditioning operation data as described in any one of claims 1 to 4.

Images