JP2023530628A - Cooling equipment control method, cooling equipment control device, computer equipment and computer readable medium - Google Patents

Abstract

The present disclosure includes the steps of determining a current outdoor temperature, and inputting synchronized historical sample data of a cooling equipment load and a preset influence factor as first input parameters into a first neural network model, and inputting the synchronous history sample data of the outdoor temperature and the predicted load of the cooling equipment on the day as the second input parameter to the second neural network to obtain the predicted indoor temperature on the day. and inputting the indoor temperature predicted for the day and a preset cooling efficiency factor as a third input parameter to a third neural network to obtain the optimum control parameters for the cooling equipment for the day; and cooling based on the optimum control parameters. and controlling operation of the equipment. The present disclosure further provides cooling equipment controllers, computer equipment and computer readable media. [Selection drawing] Fig. 5

Description

TECHNICAL FIELD The present disclosure relates to the technical field of automatic control, and in particular to a method for controlling cooling equipment, a cooling equipment control device, a computer equipment and a computer-readable medium.

In mobile communication networks, about 80% of the power consumption comes from widely distributed base stations, and the more dense the base stations, the higher the power consumption. Normally, the machine room of the base station selects different capacity air conditioners according to its type (brick building, hut, colored steel building, etc.) and the equipment load in the machine room, and the overheated equipment and It lowers the temperature of the equipment and ensures the safe operation of the equipment. Cooling of air conditioners accounts for the largest proportion of power consumption in mobile base stations, and optimizing air conditioner control algorithms will reduce power consumption and electricity charges in mobile communication base stations. It has become one of the most important directions to work for.

The present disclosure includes the steps of determining the current outdoor temperature, and inputting the synchronous history sample data of the cooling equipment load and the preset influence factor as the first input parameters into the first neural network model to obtain the predicted current day temperature of the cooling equipment. inputting the synchronous history sample data of the outdoor temperature and the predicted load of the cooling equipment on the day as a second input parameter to a second neural network to obtain the predicted indoor temperature on the day; a step of inputting an indoor temperature predicted for the day and a preset cooling efficiency factor as a third input parameter to a third neural network to obtain an optimum control parameter for the cooling equipment for the day; and controlling operation of the cooling equipment.

The present disclosure comprises a first processing module, a second processing module, and a control module, wherein the first processing module is used to determine a current outdoor temperature, and the second processing module is a cooling device. inputting the synchronous history sample data of the load and the preset influence factor into the first neural network model as a first input parameter to obtain the predicted load of the cooling equipment for the day; The predicted load of the cooling equipment on the day is input to the second neural network as a second input parameter to obtain the predicted room temperature on the day, and the predicted room temperature on the day and the preset cooling efficiency factor are input to the second neural network. input to a third neural network as three input parameters to obtain the optimum control parameters of the cooling equipment for the day, and the control module controls the operation of the cooling equipment based on the optimum control parameters; Further provided is a refrigeration equipment controller for use in:

The present disclosure includes one or more processors and a storage device storing one or more programs, and when the one or more programs are executed by the one or more processors, the Further provided is a computer apparatus in which one or more processors implement the methods of the present disclosure.

The present disclosure further provides a computer readable medium having stored thereon a computer program which, when executed by a processor, causes the processor to implement the method of the present disclosure.

1 is a schematic diagram of a cooling appliance control system according to the present disclosure; FIG. FIG. 4 is a diagram showing a flow of creating first, second, and third neural network models according to the present disclosure; FIG. 4 is a diagram showing a flow of creating first, second, and third neural network models according to the present disclosure; FIG. 4 is a diagram showing a flow of creating first, second, and third neural network models according to the present disclosure; FIG. 4 is a diagram showing a control flow of cooling equipment according to the present disclosure; 1 is a schematic diagram of a first neural network model according to the present disclosure; FIG. FIG. 3 is a schematic diagram of a second neural network model according to the present disclosure; FIG. 3 is a schematic diagram of a third neural network model according to the present disclosure; It is a figure which shows the control flow of the air conditioner which concerns on this indication. FIG. 4 is a diagram showing a control flow of a heat exchanger according to the present disclosure; FIG. 4 is a diagram showing a flow of re-determining and updating optimal control parameters for the current day for cooling equipment according to the present disclosure; 1 is a schematic diagram showing a configuration of a cooling equipment control device according to the present disclosure; FIG. 1 is a schematic diagram showing a configuration of a cooling equipment control device according to the present disclosure; FIG.

DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS Reference will now be made in detail to exemplary embodiments with reference to the drawings, which can be embodied in different forms and which is understood to be limited to the embodiments set forth herein. shouldn't. These examples are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used herein, the singular forms "one" and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. Also, when the terms "comprising" and/or "consisting of" are used herein, the presence of these features, wholes, steps, acts, elements and/or components is indicated. implied, but does not exclude the presence or addition of one or more of other features, elements, steps, acts, elements, components, and/or groups thereof.

Embodiments described herein may be described with reference to plan and/or cross-sectional views in an idealized schematic representation of this disclosure. Accordingly, exemplary illustrations may be modified based on manufacturing technology and/or tolerances. Accordingly, the embodiments are not limited to the illustrated embodiments, but include variations in configuration made based on the manufacturing steps. Thus, the illustrated exemplary regions have exemplary attributes, and the illustrated region shapes indicate specific shapes of regions of elements, but are not intended to be limiting.

Unless defined otherwise, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art. Also, terms defined in commonly used dictionaries should be construed to have the same meaning as they have in the context of the relevant technical field and this disclosure, and unless stated otherwise to limit this disclosure, ideal should not be generalized or interpreted too formally.

When expanding the machine room of a base station, operators should install energy-saving type air conditioners instead of air conditioners or interlocking control with air conditioners in order to ensure long-term and stable operating conditions for various equipment in the machine room. Consider placing a heat exchanger for The basic principle of an energy-saving heat exchanger is to use the outdoor natural environment as a source of cold heat. By reducing the temperature of the machine room by letting the heat escape, the operating time of the air conditioner is shortened and power is saved.

The general interlocking control algorithm is the traditional temperature control start/stop method, and the environment temperature is the main basis for the heat exchanger and air conditioner interlocking control. The algorithm is simple but difficult to improve. The general interlocking control process detects the indoor and outdoor temperatures in real time, and when the indoor temperature exceeds the upper limit of the operating temperature of the equipment, the heat exchanger or air conditioner will start to cool, and the heat exchanger will is satisfied (when the temperature difference between indoors and outdoors reaches a threshold), the heat exchanger is preferentially activated, otherwise the air conditioner is activated. Frequent switching of air conditioners and heat exchangers is not desirable and should be performed at intervals of half an hour or longer.

The start/stop condition parameters for the heat exchanger and the air conditioner are set separately. When it exceeds, the heat exchanger is turned on, when the room temperature falls below 25°C, it is turned off, and when the temperature difference between indoor and outdoor exceeds 8°C, the heat exchanger can be started. However, in practical use, it is usually difficult to determine the starting and stopping condition parameters, and the starting and stopping condition parameters differ depending on the region, seasonal climate, and the temperature difference between morning and evening. If so, the air conditioner will be activated frequently, resulting in an increase in power consumption. A general interlocking control algorithm only considers an external factor such as the environmental temperature, and cannot predict the startup time and the number of startups of the air conditioner.

Factors such as seasonal climatic changes, temperature changes, load fluctuations, equipment actual temperatures, etc., and combinations of changes in these factors, all affect the operating policy of cooling equipment and therefore the cooling Device control policies lack rules to follow. Assuming that the start-up temperature of the heat exchanger is 35°C and the start-up temperature of the air conditioner is 40°C, assuming that the room temperature of a certain base station rarely exceeds 35°C, the heat exchanger is normally turned on. Assume that the load requirement is met and the air conditioner does not need to be turned on, but the room temperature may exceed 40° C. during certain traffic peaks and temperature peaks. With the conventional control algorithm, it is necessary to turn on the air conditioner, but the high temperature time of 40 degrees Celsius or more is short and does not affect the safe operation of the equipment (the operating range of some base stations/transmission equipment is 40 degrees Celsius). °C for a long time, but reaching 50 °C for a short period of time), there is actually no need to turn on the air conditioner. As a result, it is possible to avoid turning on the air conditioner while ensuring the safety of the equipment, and energy saving is achieved to a certain extent.

The cooling equipment control method provided by the present disclosure can control the operation of the cooling equipment in the machine room. This method is applicable to the cooling control system shown in FIG. 1, and as shown in FIG. have equipment. An FSU is a field device, located in a machine room with cooling equipment, and equipped with a collection unit and an execution unit. The collection unit is used to collect and upload real-time data such as outdoor temperature and humidity, indoor temperature, equipment load, etc. to the cooling controller. The execution unit is used to control the operation of the cooling equipment according to the instructions of the cooling controller. The cooling appliance controller may be a cloud appliance, and includes a first neural network (NN) model, a second neural network model, a third neural network model, a historical sample data bank and a cooling appliance control policy ( For example, a Unified Management Expert (UME) staffed with a cooling control algorithm) may be selected. The UME obtains the predictive control scheme of the cooling equipment based on the data reported by the FSU and the first neural network model, the second neural network model, and the third neural network model, and sends the predictive control scheme to the FSU. good. Cooling equipment may include air conditioners and heat exchangers and may operate based on the transmitted control scheme.

During the initialization phase, the following first through tenth thresholds and various times may be preset in the cooling appliance controller.

The first threshold VHT may be, for example, 45° C., and the air conditioner is unconditionally activated when the room temperature exceeds VHT. The second threshold VLT may be, for example, 15° C. When the room temperature falls below VLT, the air conditioner is unconditionally stopped and is smaller than the first threshold VHT. The third threshold HT _AC may be, for example, 40° C., and the air conditioner may turn on when the room temperature exceeds HT _AC . The fourth threshold HT _HEE may be, for example, 35° C., and the heat exchanger may activate when the room temperature exceeds HT _HEE . A fifth threshold is used to determine whether a second hot prestart condition for the indirect heat exchanger is met. The sixth threshold LT may be, for example, 25° C., the air conditioner and the heat exchanger may stop when the room temperature falls below LT, the fourth threshold HT _HEE and the third threshold HT smaller than _AC . A seventh threshold is used to determine off-time of the cooling device. The eighth threshold is used to determine whether the indoor/outdoor temperature difference in the second high temperature pre-start condition of the direct heat exchanger is met. A ninth threshold is used to determine whether the humidity in the second hot pre-startup condition of the direct heat exchanger is met. The tenth threshold is used to determine the error between the air conditioner's actual operating parameters and the air conditioner's optimal control parameters. Regarding the maximum on-time MAXCOT of the air conditioner and the minimum off-time MINCST of the air conditioner, normally the maximum on-time MAXCOT of the air conditioner is 12 hours and the minimum off-time MINCST of the air conditioner is 0.5 hours.

During the initialization phase, a first neural network model, a second neural network model and a third neural network model are created. Hereinafter, the flow for creating the first neural network model, the second neural network model and the third neural network model will be described in detail with reference to FIG.

As shown in FIG. 2, this step of creating a first neural network model, a second neural network model and a third neural network model includes the following steps S21-S23.

In step S21, history sample data is obtained.

Sample data may include outdoor temperatures, indoor temperatures, and cooling equipment loads.

In step S21, the cooling appliance controller may obtain historical sample data from a historical sample data bank. A historical sample data bank may store a large amount of historical sample data such as daily outdoor temperature _TRout , indoor temperature _TRin , cooling equipment load L _R , and the like. The sampling period may be determined according to the degree of change in these parameters. For example, the sampling period of the outdoor temperature T _Rout may be set to 10 minutes, and the sampling period of the indoor temperature T _Rin and the cooling equipment load L _R may be set to It may be 5 minutes.

Sample data may include simulation data and sampling data. The simulation data is data obtained by simulating the operation of the cooling equipment when the room temperature is higher than the third threshold HT _AC . The sampling data is data obtained by sampling when the indoor temperature is lower than the sixth threshold LT and the actual OFF time of the cooling device is higher than the seventh threshold. That is, if the room temperature _TRin is high and the cooling equipment needs to be operated, a dummy load may be used to simulate the actual cooling equipment and record data such as _TRout , _TRin , _LR , etc. . When the indoor temperature T _Rin is low and the cooling equipment is off for long periods of time (e.g., during seasons when the outdoor temperature T _Rout is low or at night), a large amount of existing historical sample data is used as-is to speed up the collection of historical sample data. may

In step S22, an analog simulation of the history sample data is performed to calculate daily optimum control parameters for the cooling equipment.

In step S22, by computer simulation learning, create a heat distribution diagram of the environment of the machine room, heat generating equipment and cooling equipment, perform simulation calculation of history sample data, control optimal solution vector for cooling equipment on the day (that is, cooling equipment daily optimal control parameters), and save the daily optimal control parameters of the cooling equipment as sample data tags.

Based on simulation results and daily experience, it is not desirable to turn on the air conditioner frequently. In other words, if an air conditioner has two valid values in one on time/on time (T _moment /T _hours ) tag group, the optimal control parameters for the air conditioner on the day are It means that it is turned on every time at the on-time T _moment , and the operating time is the value of the corresponding on-time T _hours .

In step S23, a first neural network model, a second neural network model and a third neural network model are created based on the historical sample data and the daily optimum control parameters of the cooling equipment.

At step S23, a first neural network model, a second neural network model and a third neural network model are created in order.

As shown in FIG. 3, the cooling appliance control method of the present disclosure performs analog simulation of historical sample data to calculate daily optimal control parameters for the cooling appliance (i.e., step S22), after the step of: Before the step of creating a first neural network model, a second neural network model and a third neural network model based on the historical sample data and the daily optimal control parameters of the cooling equipment (i.e. step S23), step S22'. ~S23' may be further included.

In step S22', a normalization process is performed on the historical sample data and the daily optimum control parameters of the cooling equipment.

Based on the following equations, a normalization process may be performed on the historical sample data and the cooling equipment's optimal daily control parameters to keep the data between the range (0,1).

Here, X _real is the true value of the real sample, X ^* is the normalized data, X _max is the maximum value or upper limit of the corresponding type of data sample, and X _min is , is the minimum or lower bound of the corresponding kind of data samples.

For the outdoor temperature T _Rout and the indoor temperature T _Rin , X _max may have an upper limit of 100° C. and X _min may have a lower limit of −40° C., so that the normalized value of each actual temperature X _real X ^* =(X _real −X _min )/(X _max −X _min )=(X _real +40)/140. For the cooling equipment load L _R , X _max may be set as the full cooling equipment load and X _min may be set to 0, so the X _real normalized value of the cooling equipment load L _R is X ^* =X _real / _Xmax .

For air conditioner on-time T _moment-AC and heat exchanger on-time T _moment-HEE (which may be in hh:mm:ss format), X _max is an upper limit of 1440 (i.e., one day is 24 * 60 = 1440 minutes), and X _min may be set to 0, so the normalized value of T _{moment - AC} and T _{moment - HEE} is X ^* = (hh * 60 + mm) / 1440 be. For air conditioner on-time T _hours-AC and heat exchanger on-time T _hours-HEE , X _max may be set as an upper limit value of 24 (i.e., there are 24 hours in a day), and X _min is: Since it can be 0, the normalized value X* of T _hours-AC and T _hours-HEE is X ^* =X _real /24.

In step S23', a learning sample data set including a learning set, a verification set and a test set is created based on the normalized data.

In step S23', a training set, a validation set and a test set may be created based on a sample ratio of 6:2:2.

Therefore, the step of creating a first neural network model, a second neural network model, and a third neural network model based on the historical sample data and the daily optimal control parameters of the cooling equipment (i.e., step S23) includes learning sample data Based on the set, creating a first neural network model, a second neural network model and a third neural network model may be included.

Hereinafter, the flow for creating the first neural network model, the second neural network model and the third neural network model will be described in detail with reference to FIG.

As shown in FIG. 4, creating a first neural network model, a second neural network model and a third neural network model based on the historical sample data and the daily optimal control parameters of the cooling appliance (i.e., step 23). may include steps S231-S233.

In step S231, a first neural network model is created using the synchronous history sample data of the cooling equipment load and the preset influence factor as the first input parameters and the history sample data of the cooling equipment load of the day as the first output parameter. .

The influence factor may include one or any combination of a holiday influence factor F _holiday , a tide influence factor F _tide , a regional event factor F _event . The value ranges of the holiday influence factor F _holiday , the ebb and flow influence factor F _tide and the regional event factor F _event are all (0, 1) and may be determined manually based on experience. For example, in residential areas, the holiday influence factor F _holiday on normal working days may be 0, the holiday influence factor F _holiday on Saturdays and Sundays may be 0.1, and the Spring Festival long holiday may be _0.25 . In the case of an industrial park, the ebb and tide effect factor F _tide during working hours may be 0.5, and the ebb and tide effect factor F _tide during overtime hours may be 0.7. The time period ebb and flow influence factor F _tide may be 0.3. For some areas, the normal area event factor F _event may be 0, the area event factor F _event with commercial operations may be 0.1, and the area event factor for gatherings F _event may be 0.2 and may be the regional event factor F _event 0.3 for the concert.

In step S232, a second neural network model is created using the synchronous history sample data of the outdoor temperature and the history sample data of the load of the cooling equipment on the day as the second input parameters, and the history sample data of the indoor temperature on the day as the second output parameter. do.

In step S233, the history sample data of the indoor temperature of the day and the preset cooling efficiency factor are set as the third input parameters, and the history sample data of the optimum control parameters of the cooling equipment for the day are set as the third output parameter, and the third neural network model to create

The optimal control parameters are the ON time T _moment and the ON time T _hours , that is, the ON time T _moment-AC of the air conditioner, the ON time T _moment-TEE of the heat exchanger, the ON time T _hours-AC of the air conditioner, and the heat exchange It may include the on-time T _hours-TEE of the device.

The cooling efficiency factors may include a heat exchange cooling efficiency factor F _eff1 and an air conditioning cooling efficiency factor F _eff2 . If the environment of the machine room is fixed, both the heat exchange cooling efficiency factor F _eff1 and the air conditioning cooling efficiency factor F _eff2 are constants. , or move the spatial position, etc.), the heat exchange cooling efficiency factor F _eff1 and the air conditioning cooling efficiency factor F _eff2 need to be adjusted to new constants.

Suppose _{that the} 24 _sets of T _{moment /} T _hours data for a heat exchanger have two valid values, e.g. is 0.10, and none of the 12 sets of T _moment /T _hours for an air conditioner have valid values. After converting the on-time T _moment back to hh:mm:ss format and the on-time T _hours back to standard time, the optimal control parameters for the day of the chiller mean:

(1) The heat exchanger was pre-on/operated twice during the day.

(2) The first ON time of the heat exchanger is 10:48 (i.e. 0.45*24=10.8=10:48) and 1.2 hours (i.e. 0.05*24=1 .2) working, ie the working time interval is 10:48-12:00 (ie 0.45*24+0.05*24=12=12:00);

(3) The second ON time of the heat exchanger is 14:24 (i.e. 0.60*24=14.4=14:24), 2.4 hours (i.e. 0.10*24=2.4 ), that is, the operating time interval is from 14:24 to 16:48 (ie, 0.60*24+0.10*24=16.8=16:48).

(4) The air conditioner did not work on the day.

After being learned and optimized, the first neural network model, the second neural network model and the third neural network model can be deployed according to the actual working environment. Any of the three neural network models may be deployed in the UME to fully utilize the cloud's massive computing resources to achieve real-time or online learning. If necessary, the three neural network models may be arranged on the edge side (for example, FSU) by adding calculation rods or the like.

FIG. 5 is a diagram showing the control flow of the cooling device according to the present disclosure.

As shown in FIG. 5, the cooling appliance control method according to the present disclosure may be used to control the operation of the cooling appliance and may include steps S11-S15.

In step S11, the current outdoor temperature is determined.

Based on the predicted temperature and the detected outdoor temperature, calculate the current outdoor temperature T _Rout by weighting, that is, first determine the outdoor temperature within a preset time before the current time, and then The current outdoor temperature T _Rout may be determined based on the outdoor temperature within a preset time before the current time, the predicted temperature for the current day, and preset first and second weighted values. Usually, the preset time may be one hour, and the today's predicted temperature may be the temperature predicted by the weather forecast for that day. For example, outdoor temperature T _Rout = weather forecast temperature * 0.8 + outdoor temperature actually measured one hour ago * 0.2.

The FSU may collect data such as indoor/outdoor temperature, humidity, and cooling equipment load and upload them to the UME.

In step S12, the synchronous history sample data of the cooling equipment load and the preset influence factor are input to the first neural network model as the first input parameters to obtain the predicted load of the cooling equipment for the day.

Synchronization refers to the same period in the past, and for example, both the same time on the same day last year and the same time on the same day last year may be synchronized with the time on the same day.

In step S12, as shown in FIG. 6A, the synchronous history sample data L _N of the cooling equipment load, the holiday influence factor F _holiday , the ebb and flow tide influence factor F _tide , and the regional event factor F _event are input into the first neural network model, The predicted load L _R of the cooling equipment for the day is obtained and used as the output value of the first neural network model.

In step S13, the synchronous history sample data of the outdoor temperature and the predicted load of the cooling equipment on the day are input to the second neural network as second input parameters to obtain the indoor temperature predicted on the day.

In step S13, as shown in FIG. 6B, the synchronous history sample data T _Rout of the outdoor temperature and the load L _R (that is, the output value of the first neural network) predicted for the day of the cooling equipment are input to the second neural network model. Then, the indoor temperature _TRin predicted for the day is obtained and used as the output value of the second neural network model.

In step S14, the indoor temperature predicted for the day and the preset cooling efficiency factor are input as third input parameters to the third neural network to obtain the optimum control parameters for the cooling equipment for the day.

In step S14, as shown in FIG. 6C, the indoor temperature T _Rin (that is, the output value of the second neural network) predicted for the day, the heat exchange cooling efficiency factor F _eff1 and the air conditioning cooling efficiency factor F _eff2 are applied to the third neural network. Input into the model to determine the optimal control parameters for the day of the air conditioner (that is, the air conditioner ON time T _moment-AC and the air conditioner ON time T _hours-AC ) and the heat exchanger optimal control parameters for the day (that is, Obtain the heat exchanger on-time T _moment-TEE and the heat exchanger on-time T _hours-TEE ).

In actual use, the on-time T _moment can be returned to the hh:mm:ss format, and the on-time T _hours can be returned to standard time (eg, xx hours).

In steps S12 to S14, the UME sequentially operates the first neural network model, the second neural network model and the third neural network model to output the optimum control parameters of the cooling equipment for that day.

The optimal control parameters of the air conditioner may include the on-time T _moment-AC of the air conditioner and the on-time T _hours-AC of the air conditioner, and the optimal control parameter of the heat exchanger is the on-time T _moment-AC of the heat exchanger. _TEE and heat exchanger on-time T _hours-TEE may also be included.

The optimal control parameters are the on-time T _moment-AC and on-time T _hours-AC of up to 12 sets of air conditioners every day, and the on-time T _moment-TEE of up to 24 sets of heat exchangers and heat exchangers. It may also include the on-time T _hours-TEE .

In step S15, the operation of the cooling equipment is controlled based on the optimum control parameters.

In step S15, the operation of the air conditioner is controlled based on the optimum control parameters of the air conditioner, and the operation of the heat exchanger is controlled based on the optimum control parameters of the heat exchanger.

According to the cooling device control method according to the present disclosure, a neural network model is used to combine parameters such as the current outdoor temperature, the cooling device load synchronization history sample data, the influence factor, and the cooling efficiency factor to It realizes prediction and interlocking control of the control scheme of the heat exchanger. It provides control, optimizes operating efficiency, and reduces power consumption. In addition, by combining past data and current measured data, the influence factor of special incidents and the influence factor of cooling efficiency of cooling equipment are taken into account to make the predicted control scheme more accurate. It is possible to change the environment, and the range of application is improving.

FIG. 7 is a diagram showing a control flow of an air conditioner according to the present disclosure.

As shown in FIG. 7, the air conditioner control flow according to the present disclosure includes steps S31 to S39.

In step S31, if the current indoor temperature is greater than the first threshold value VHT, step S36 is executed; otherwise, step S32 is executed.

In step S31, if the current indoor temperature is higher than VHT, it means that the current indoor temperature is too high, and it may be further determined whether the operation of the air conditioner has timed out (that is, step S36 is executed). ). If the current indoor temperature is below VHT, it may be further determined whether the current indoor temperature is too low (that is, step S32 is executed).

In step S32, if the current indoor temperature is lower than the second threshold value VLT, step S39 is executed; otherwise, step S33 is executed.

In step S32, if the current indoor temperature is less than the second threshold VLT, it means that the current indoor temperature is too low, and the air conditioner may be turned off due to low temperature abnormality (that is, step S39 ). If the current room temperature is greater than or equal to the second threshold VLT, it means that the current room temperature is not turned off due to a high temperature abnormality and is not turned off due to a low temperature abnormality. Further determination may be made (that is, step S33 is executed).

In step S33, if the first high-temperature pre-start condition is satisfied, step S34 is executed; otherwise, the process returns to step S31.

In step S33, if the current indoor temperature is equal to or lower than the first threshold value VHT and equal to or higher than the second threshold value VLT, and the first high-temperature pre-activation condition is satisfied, the air conditioner is based on the optimum control parameters for the current day. If the operation is controlled (that is, step S34 is executed) and the first high temperature pre-start condition is not satisfied, the process returns to step S31.

The first high-temperature pre-activation condition is that the air conditioner on time T _moment-AC is reached, the current indoor temperature is greater than the third threshold HT _AC , and the actual off time of the air conditioner is the shortest off time of the air conditioner. Exceeding MINCST may also be included.

In step S34, the maximum operating time T _on-max of the air conditioner is set to the minimum value of the on-time T _hours-AC of the air conditioner and the maximum on-time MAXCOT of the air conditioner.

In step S34, in order to ensure reliability and safety in the operation of the air conditioner, the minimum value of the on-time T _hours-AC of the air conditioner and the maximum on-time MAXCOT of the air conditioner is taken. may be used as control parameters for actual control.

In step S35, the air conditioner is activated and step S38 is executed.

In step S35, after controlling the start-up of the air conditioner, recording of the actual ON time T _on-AC of the air conditioner is started, the actual OFF time T _off-AC of the air conditioner is cleared, and step S38 is executed.

In step S36, if the actual off-time Toff _-AC of the air conditioner is greater than the minimum off-time MINCST of the air conditioner, step S37 is executed; otherwise, the process returns to step S31.

In step S36, if the current indoor temperature is greater than the first threshold value VHT and the current actual off time Toff _-AC of the air conditioner is greater than the shortest air conditioner off time MINCST, it is determined that the high temperature abnormality start condition is satisfied. It means that the air conditioner high temperature abnormal start operation is executed (that is, step 37 is executed). If the current actual off-time Toff _{-AC of} the air conditioner is less than or equal to the shortest off-time MINCST of the air conditioner, the process returns to step S31.

In step S37, the maximum operating time T _on-max of the air conditioner is set as the maximum on time MAXCOT of the air conditioner, and step S35 is executed.

In step S37, when the air conditioner starts abnormally at high temperature, the operating time of the air conditioner is controlled directly based on the preset maximum on-time MAXCOT of the air conditioner.

In step S38, if the actual on-time T _on-AC of the air conditioner is greater than or equal to the maximum operating time T _on-max of the air conditioner, execute step S39; otherwise, keep the current state of the air conditioner. .

After the air conditioner is started, recording of the actual ON time T _on-AC of the air conditioner is started, and if the actual ON time T _on-AC of the air conditioner is equal to or greater than the maximum operating time T _on-max of the air conditioner, the air conditioner stop the air conditioner, otherwise the air conditioner keeps its current state.

In step S39, the air conditioner is stopped and the process returns to step S31.

In step S39, after controlling the stop of the air conditioner, the recording of the actual off time T _off-AC of the air conditioner is started, the actual on time T _on-AC of the air conditioner is cleared, and then the process returns to step S31. Continue to detect room temperature.

The air conditioner control flow may further include stopping the air conditioner if the current indoor temperature is less than a sixth threshold LT.

As can be seen from the above steps S31 to S39, based on applying the prediction scheme output from the third neural network model, by combining preset air conditioner and heat exchanger start/stop policy algorithms, , safe operation of air conditioners and heat exchangers is guaranteed even when there is an abnormality in the prediction of the neural network model. When the actual room temperature exceeds the first threshold value VHT, the air conditioner may start due to the high temperature abnormality. When the actual room temperature falls below the second threshold VLT, the air conditioner may be turned off due to low temperature anomaly. When the on-time T _moment-AC of the air conditioner is reached, and the actual room temperature exceeds the third threshold HT _AC , and the operation interval exceeds the minimum off-time MINCST, the air conditioner outputs That is, the operation is activated when the air conditioner on-time T _moment-AC is reached, and the operation duration is the air conditioner on-time T _hours-AC .

FIG. 8 is a diagram showing a control flow of a heat exchanger according to the present disclosure;

As shown in FIG. 8, the control flow of the heat exchanger according to the present disclosure includes steps S41-S44.

In step S41, if the second hot pre-startup condition is met, execute step S42, otherwise the heat exchanger keeps the current state.

The heat exchanger may include a direct heat exchanger and an indirect heat exchanger, the direct heat exchanger may be equipped with a fresh air system, and the indirect heat exchanger may be a heat pipe equipment. , HPE).

When the heat exchanger is an indirect heat exchanger, the second high temperature pre-startup condition is that the heat exchanger's on time T _moment-TEE is reached and the current room temperature is greater than the fourth threshold HT _HEE , and The difference value between the current indoor temperature and the outdoor temperature may be greater than a fifth threshold.

When the heat exchanger is a direct heat exchanger, the second hot pre-start conditions include either:

(1) The heat exchanger ON time T _moment-TEE is reached, the current indoor temperature is greater than the fourth threshold HT _HEE , and the difference value between the current indoor temperature and the outdoor temperature is the eighth threshold greater than The eighth threshold may be greater than the fifth threshold, that is, in the second high-temperature pre-start condition, the indoor-outdoor temperature difference requirement of the direct heat exchanger is equal to the indoor-outdoor temperature difference of the indirect heat exchanger. Greater than the requirement, for example, the fifth threshold may be 6°C and the eighth threshold may be 10°C.

(2) The heat exchanger on-time T _moment-TEE is reached, the current indoor temperature is greater than the fourth threshold HT _HEE , and the difference between the current indoor temperature and the outdoor temperature is the eighth threshold is greater and the current indoor humidity is less than or equal to the ninth threshold. That is, the second hot pre-start conditions of the direct heat exchanger may include temperature and humidity conditions, for example, the ninth threshold may be 90%.

At step S42, the heat exchanger is activated.

In step S42, after controlling the start-up of the heat exchanger, the recording of the actual on-time T _on-HEE of the heat exchanger is started, and the actual off-time T _off-HEE of the heat exchanger is cleared.

In step S43, if the actual on-time T _on-HEE of the heat exchanger is greater than or equal to the on-time T _hours-HEE of the heat exchanger, execute step S44; Hold.

At step S44, the heat exchanger is stopped.

In step S44, after controlling the shutdown of the heat exchanger, the recording of the actual OFF time T _off-HEE of the heat exchanger is started, and the actual ON time T _on-HEE of the heat exchanger is cleared.

The heat exchanger control flow may further include shutting down the heat exchanger if the current room temperature is less than a sixth threshold LT.

The air conditioner and the indirect heat exchanger may operate simultaneously, but the operation of the air conditioner and the direct heat exchanger is exclusive, that is, both the air conditioner and the direct heat exchanger operate simultaneously. must not. Also, in the event of a delay or fire alarm, it is necessary to immediately stop the operation of the direct heat exchanger and close the damper to ensure safety.

If the heat exchanger is a direct heat exchanger, the turn-on time of the air conditioner and the turn-on time of the heat exchanger are different. Therefore, the cooling device control method of the present disclosure may further include stopping the heat exchanger when the air conditioner is turned on, and stopping the air conditioner when the heat exchanger is turned on.

The control algorithms for air conditioners and heat exchangers may be operated in the UME cloud, and if necessary, the control algorithms for air conditioners and heat exchangers may be copied to the FSU and executed locally. . In this case, the UME needs to pre-send to the FSU the cooling control scheme predicted by the third neural network.

Note that the air conditioner control and the heat exchanger control may be executed in parallel, and steps S11 to S14 shown in FIG. .

FIG. 9 is a diagram showing a re-determining and updating flow of the optimum control parameters for the current day of the cooling equipment according to the present disclosure.

As shown in FIG. 9, the method for controlling cooling equipment of the present disclosure further includes steps S51 to S53 after the step of controlling the operation of the cooling equipment based on the optimum control parameters (that is, step S15 shown in FIG. 5). may contain.

In step S51, if the error between the actual operation parameter of the air conditioner for the day and the optimum control parameter of the air conditioner for the day exceeds a tenth threshold, step S52 is executed; exit.

In step S52, the optimum control parameters of the air conditioner for the current day are determined again.

The specific implementation method of step S52 is the same as steps 12 to S14, and will not be repeated here.

In step S53, the learning sample data set is updated based on the re-determined optimal control parameters of the air conditioner for that day.

For example, if the error between the actual turn-on time of the air conditioner on the day and the turn-on time of the air conditioner in the optimum control parameters of the air conditioner on the day exceeds 10 minutes, the ability to quickly adapt the cooling control policy and predictive control In order to improve real-time performance and accuracy, it is necessary to re-predict the optimum control parameters of the air conditioner for the day and update the learning sample data set based on the re-predicted optimum control parameters of the air conditioner for the day.

Neural network models may be deployed and run in the cloud, and if the external parameters do not change constantly, these models will constantly improve their prediction accuracy and adapt to abnormal situations such as changes in the machine room environment to learn and In addition, real-time or online learning can occur constantly so that adjustments can be made.

Air conditioners and heat exchangers are fault-tolerant, so the cooling device control method of the present disclosure can be used even if the currently operating cooling device has a fault and the other cooling device is normal. If there is, turn off the failed cooling device and turn on the normal cooling device, and if both of the two cooling devices currently running are faulty, when the fault is cleared, the fault will occur. and activating the cleared cooling device. That is, if a cooling device that is currently on is faulty, turn off the faulty cooling device, turn on the good cooling device, turn on the cooling device again after the fault is cleared, and turn it on again. shut down one cooling device. By starting and operating the air conditioner and the heat exchanger as backups in the event of a failure, it is possible to avoid the risk of the machine room becoming abnormally hot.

In the cooling device control method of the present disclosure, in the process of controlling the operation of the cooling device based on the optimum control parameters (that is, step S15 shown in FIG. 5), the current room temperature is lower than the sixth threshold LT, and if the actual off time of the cooling equipment is greater than the seventh threshold, learning a second neural network model based on currently acquired sample data including the outdoor temperature, the indoor temperature, and the cooling equipment load. It may contain further. That is, if environmental conditions are favorable (e.g., Fast Ethernet connection between FSU and Cloud UME, and Cloud UME computing resources are sufficient), real-time or online model training may be supported. good. When the temperature in the machine room is low and the cooling equipment is not in operation for a long time (for example, in the cool season or at night when the temperature is low), based on data such as outdoor temperature, equipment load, and indoor temperature collected in real time, A second neural network model may be trained in real-time and online.

In the method for controlling a cooling device of the present disclosure, after controlling the operation of the cooling device based on the optimal control parameters (that is, step 15 shown in FIG. 5), the first neural network model and the first neural network model are based on the learning sample data set. The method may further include adding the acquired sample data of the day and the operating parameters of the cooling equipment of the day to a training sample data set so that the three-neural network model can be trained. By adding the sample data of the day and the actual cooling control results to the big data set, the learning set and test set are enriched, and the first and third neural network models are trained online to improve the prediction accuracy of the model. can be enhanced.

If the communication network is interrupted, the FSU cannot communicate with the UME, and the FSU may automatically operate the built-in temperature start/stop control algorithm to realize the control of the cooling equipment, and the UME A pre-transmitted cooling control plan may be received and stored to locally operate a cooling interlock control algorithm copied from the UME.

Therefore, the control method of the cooling device of the present disclosure, after creating the first neural network model, the second neural network model, and the third neural network model in the initialization stage, when the communication between the FSU and the UME fails, A first neural network model, a second neural network model, and a third neural network model are placed in the FSU so as to determine the optimal control parameters for the current day of the equipment and control the operation of the cooling equipment based on said optimal control parameters. It may further include:

One application scene of the present disclosure is the machine room of the base stations whose heat value of the communication equipment in the machine room is less than 10 KW, usually the base station machine room of the operator's data, transmission and switching. Conventional cooling equipment for the machine room consisted of only one air conditioner, but in order to reduce the power consumption of the air conditioner, the external environment, heat generation amount, and installation conditions of the machine room of the base station were considered. By adding an indirect heat exchanger such as a smart heat pipe system (HPE), the machine room solution scheme is realized by interlocking control of the air conditioner and the heat pipe system. The machine does not need to be cooled to use heat pipe technology, and the temperature difference between indoors and outdoors is basically kept around 6 degrees, so it can be used more than 90% of the time throughout the year. In addition, the power consumption of the parts is much lower than that of air conditioning by conventional compressors, and is only about 1/5 of that of conventional air conditioner systems, so that the power consumption of air conditioners can be greatly saved.

Under normal circumstances, whether the machine room is new or expanded, the working environment of the machine room is fully considered so that the appropriate heat exchanger can be selected. Indirect heat exchangers, such as heat pipes and heat exchangers, can separate the inside and outside environments, and have a wide range of applications, but the initial investment cost is high. If the air quality is good (no smoke, no corrosive gas contamination), the temperature and humidity are low, and the user has great ability to maintain it regularly, a fresh air system may be selected as the direct heat exchanger. . Therefore, another application scene of the present disclosure is a base station machine room with interlocking cooling using a fresh air system and an air conditioner.

According to the cooling equipment control scheme according to the present disclosure, based on big data technology and neural network technology, the current indoor and outdoor temperature and humidity, system load, etc. are fully considered, load prediction, weather forecast, synchronization history sample Calculations are performed using a neural network based on data, etc., to predict the cooling equipment load and room temperature in advance, output an optimal interlocking control plan for the cooling equipment for the day, and combine it with conventional control rules and policies. Therefore, the predictable active control of the air conditioners and heat exchangers in the machine room can be realized, and the objectives of optimization control, energy saving and resource saving can be achieved.

Due to the predictable and active cooperation of heat exchangers and air conditioners, the operating time and number of starts of air conditioners are significantly reduced. In addition, the operating temperature of the equipment in the machine room can be raised to a controllable safe range of 30-40° C., and the power consumption of the cooling equipment can be further reduced. Preliminary estimates show that the interlocking control scheme that actively predicts air conditioners and heat exchangers will reduce the power consumption of communication base stations by nearly 10,000 kWh each year, compared to the cooling method that uses only air conditioners. The amount can be reduced by 40%, and if calculated at the rate of 10% of the base station using 5 million kwh, the electricity bill will be reduced by 5 billion yuan and the carbon emission will be reduced by 1.35 million tons every year, and the benefits to the economy and society will be remarkable. is.

Based on the same technical idea, the present disclosure further provides a cooling equipment control device.

10 and 11 are schematic diagrams showing the configuration of the cooling equipment control device according to the present disclosure.

As shown in FIG. 10, the cooling equipment control device according to the present disclosure includes a first processing module 101, a second processing module 102 and a control module 103. As shown in FIG.

A first processing module 101 is used to determine the current outdoor temperature.

the second processing module 102 inputs the synchronous historical sample data of the cooling equipment load and the preset influence factor as the first input parameters into the first neural network model to obtain the today's predicted load of the cooling equipment; Synchronous history sample data of outdoor temperature and the predicted load of the cooling equipment on the day are input to a second neural network as a second input parameter to obtain the predicted indoor temperature on the day; Inputting a preset cooling efficiency factor as a third input parameter into a third neural network to obtain the day's optimum control parameters of the cooling equipment.

A control module 103 is used to control the operation of the cooling equipment based on the optimal control parameters.

As shown in FIG. 11 , the cooling appliance control apparatus of the present disclosure may further include a model creation module 104 .

Model creation module 104 is used to create the first, second and third neural network models in an initialization phase. The modeling module 104 acquires historical sample data including outdoor temperature, indoor temperature, and cooling equipment load, performs analog simulation of the historical sample data, and calculates the daily optimal control parameters of the cooling equipment; and creating a first neural network model, a second neural network model and a third neural network model based on the historical sample data and the daily optimal control parameters of the cooling appliance.

After the model creation module 104 performs an analog simulation of the historical sample data and calculates daily optimal control parameters of the cooling device, based on the historical sample data and the daily optimal control parameters of the cooling device, normalizing the historical sample data and the daily optimal control parameters of the cooling appliance before creating a first neural network model, a second neural network model and a third neural network model; and may be further used to create a training sample data set including a training set, a validation set and a test set based on the data obtained from the

A model creation module 104 may be used to create a first neural network model, a second neural network model, and a third neural network model based on the training sample data set.

The model generation module 104 uses the synchronous history sample data of the cooling equipment load and the preset influence factor as first input parameters, and the current day load history sample data of the cooling equipment as a first output parameter, to create a first neural network. creating a model, using the synchronous history sample data of the outdoor temperature and the history sample data of the load of the cooling equipment on the day as a second input parameter, and the history sample data of the indoor temperature on the day as a second output parameter, a second neural Creating a network model, using the history sample data of the room temperature of the day and the preset cooling efficiency factor as a third input parameter, and the history sample data of the optimum control parameter of the cooling equipment for the day as a third output parameter. , and to create a third neural network model.

Sample data may include simulation data and sampling data. The simulation data is data obtained by simulating the operation of the cooling equipment when the room temperature is higher than a preset third threshold. The sampling data is data obtained by sampling when the room temperature is lower than a preset sixth threshold and the actual off time of the cooling device is higher than a preset seventh threshold.

The first processing module 101 determines the outdoor temperature within a preset time before the current time, and determines the outdoor temperature within the preset time before the current time, the predicted temperature of the day and the preset temperature. It may be used to determine the current outdoor temperature based on the set first and second weighting values.

Optimal control parameters may include on-time and on-time. The influence factor may include one or any combination of a holiday influence factor, a tide influence factor, a local event factor.

The control module 103 determines that the current indoor temperature is equal to or lower than a preset first threshold and equal to or higher than a second threshold lower than the preset first threshold, and a first high temperature preliminary If the start condition is satisfied, setting the maximum operating time of the air conditioner to the minimum value of the on time of the air conditioner and the preset maximum on time of the air conditioner, starting the air conditioner, and the actual on time of the air conditioner is greater than or equal to the maximum operating time of the air conditioner, the air conditioner may be stopped.

The first high-temperature pre-activation condition is that the air conditioner reaches the ON time of the air conditioner, the current indoor temperature is greater than a preset third threshold value, and the actual off time of the air conditioner is preset. It may include greater than the minimum off-time.

In the process of controlling the operation of the cooling equipment based on the optimum control parameters, the control module 103 controls whether the current indoor temperature is greater than the first threshold and the actual off time of the air conditioner is the shortest of the air conditioner. if greater than the off time, setting the maximum operating time of the air conditioner to the maximum on time of the air conditioner and activating the air conditioner; and/or if the current indoor temperature is less than the second threshold, the air conditioning It may also be used to stop the machine.

The control module 103 activates a heat exchanger if a second high temperature pre-activation condition is satisfied, and activates the heat exchanger if the actual ON time of the heat exchanger is greater than or equal to the ON time of the heat exchanger. It may also be used for stopping.

When the heat exchanger is an indirect heat exchanger, the second high-temperature pre-start condition is that the heat exchanger's on-time is reached, and the current room temperature is greater than a preset fourth threshold, and The difference value between the current indoor temperature and the outdoor temperature may be greater than a preset fifth threshold.

When the heat exchanger is a direct heat exchanger, a second high temperature pre-startup condition is that the on-time of the heat exchanger is reached, and the current room temperature is greater than a preset fourth threshold, and The difference value between the current indoor temperature and the outdoor temperature is greater than a preset eighth threshold that is greater than the fifth threshold, and the heat exchanger is turned on and the current indoor temperature is greater than the preset fourth threshold, and the difference between the current indoor temperature and the outdoor temperature is greater than the preset eighth threshold, and the current indoor humidity is the preset ninth being less than or equal to a threshold.

When the heat exchanger is a direct heat exchanger, the turn-on time of the air conditioner and the turn-on time of the heat exchanger are different. The control module 103 may further be used to turn on the air conditioner to turn off the heat exchanger and turn on the heat exchanger to turn off the air conditioner.

After controlling the operation of the cooling equipment based on the optimum control parameters, the control module 103 controls the error between the actual operation parameters of the air conditioner on the day and the optimum control parameters of the air conditioner on the day. if it exceeds the 10th threshold, instruct the second processing module 102 to re-determine the day's optimum control parameters of the air conditioner; It may also be used to update the sample data set.

If the currently operating cooling device has a failure and the other cooling device is normal, the control module 103 stops the failed cooling device and turns on the normal cooling device; It may further be used to wake up the cooling device whose fault has been cleared when both of the two cooling devices operating have failed.

The second processing module 102, if the current indoor temperature is less than a preset sixth threshold and the actual off time of the cooling device is greater than a preset seventh threshold, the currently obtained It may further be used to train the second neural network model based on sample data including outdoor temperature, indoor temperature and cooling equipment load.

The present disclosure includes one or more processors and a storage device storing one or more programs, and when the one or more programs are executed by the one or more processors, the Further provided is computer equipment in which one or more processors implement the cooling equipment control method according to the present disclosure.

The present disclosure further provides a computer readable medium on which a computer program is stored, and upon execution of the computer program by a processor, the method for controlling a cooling appliance according to the present disclosure is implemented by the processor.

Those skilled in the art will understand that the steps in the methods disclosed above and the functional modules/units in all or part of the apparatus can be implemented as software, firmware, hardware and suitable combinations thereof. In a hardware embodiment, the division between functional modules/units referred to in the above description does not necessarily correspond to the division of physical components. For example, one physical component may have multiple functions, or one function or step may be performed by cooperation of several physical components. some of the physical components or all of the physical components may be implemented as software executed by a processor (CPU, digital signal processor or microprocessor) or as hardware; Alternatively, it may be implemented as an integrated circuit (eg, an application specific integrated circuit). Such software may be distributed in computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As is known to those of skill in the art, the term computer storage media includes both volatile and volatile media implemented in any method or technology for storage of information (e.g., computer readable instructions, data structures, program modules, or other data). Non-volatile, removable and non-removable media are also included. Computer storage media may be RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disc (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage. It includes, but is not limited to, a device or any other computer-accessible medium that can be used to store desired information. Also, as is well known to those skilled in the art, communication media typically includes computer readable instructions, data structures, program modules or other data in a modulated data signal, such as a carrier wave or other transport mechanism, and can be any information delivery medium. may include

Although exemplary embodiments are disclosed herein and specific terms are used, they are to be used and construed in a general and descriptive sense only and are intended to be limiting. not a thing In some examples, features, characteristics and/or elements described in connection with a particular embodiment may be used alone or described in connection with other embodiments, unless stated otherwise. One skilled in the art will appreciate that any combination of features, properties and/or elements may be used. It will therefore be appreciated by those skilled in the art that changes may be made in various forms and details without departing from the scope of the present disclosure as indicated by the appended claims.

Claims (19)

determining the current outdoor temperature;
inputting the synchronized historical sample data of the cooling equipment load and the preset influence factor as a first input parameter into a first neural network model to obtain the today's predicted load of the cooling equipment;
inputting the synchronous history sample data of the outdoor temperature and the predicted load of the cooling equipment on the day as a second input parameter to a second neural network to obtain the predicted indoor temperature on the day;
inputting the indoor temperature predicted for the day and a preset cooling efficiency factor as a third input parameter into a third neural network to obtain the optimum control parameters for the cooling equipment for the day;
and controlling the operation of the cooling equipment based on the optimum control parameters. Before the step of determining the current outdoor temperature,
obtaining historical sample data including outdoor temperature, indoor temperature and cooling equipment load;
performing an analog simulation of the historical sample data to calculate daily optimal control parameters for the cooling equipment;
and creating the first, second and third neural network models based on the historical sample data and daily optimal control parameters of the cooling equipment. Method. After the step of performing an analog simulation of the historical sample data to calculate daily optimal control parameters of the cooling device, the first neural network based on the historical sample data and the daily optimal control parameters of the cooling device Before the step of creating the network model, the second neural network model and the third neural network model,
performing a normalization process on the historical sample data and the daily optimal control parameters of the cooling equipment;
and creating a learning sample data set including a training set, a validation set and a test set based on the normalized data,
creating the first neural network model, the second neural network model and the third neural network model based on the historical sample data and the daily optimal control parameters of the cooling equipment;
3. The method of claim 2, further comprising creating a first neural network model, a second neural network model and a third neural network model based on the training sample data set. creating a first neural network model, a second neural network model, and a third neural network model based on the historical sample data and the daily optimal control parameters of the cooling equipment;
creating a first neural network model using the synchronous history sample data of the cooling equipment load and the preset influence factor as a first input parameter and the historical sample data of the current day load of the cooling equipment as a first output parameter; ,
Generating a second neural network model using the synchronous history sample data of the outdoor temperature and the history sample data of the load of the cooling equipment on the day as a second input parameter, and the history sample data of the indoor temperature on the day as a second output parameter. ,
A third neural network model is created by using the history sample data of the room temperature of the day and the preset cooling efficiency factor as a third input parameter and the history sample data of the optimal control parameter of the cooling equipment for the day as a third output parameter. 3. The method of claim 2, comprising: The sample data includes simulation data and sampling data,
The simulation data is data obtained by simulating the operation of the cooling equipment when the room temperature is higher than a preset third threshold,
The sampling data is data obtained by sampling when the indoor temperature is lower than a preset sixth threshold and the actual off time of the cooling device is higher than a preset seventh threshold. A method according to claim 2. Determining the current outdoor temperature includes:
determining the outdoor temperature within a preset time period prior to the current time;
determining a current outdoor temperature based on an outdoor temperature within a preset time period before the current time, a predicted temperature for the current day, and preset first and second weighted values. Item 1. The method according to item 1. The cooling device includes an air conditioner and a heat exchanger,
The optimal control parameters include an air conditioner ON time, an air conditioner ON time, a heat exchanger ON time, and a heat exchanger ON time,
7. A method according to any one of claims 1 to 6, wherein the influence factor comprises one or any combination of a holiday influence factor, an ebb and flow influence factor, a local event factor. The step of controlling operation of the cooling equipment based on the optimal control parameters includes:
The current indoor temperature is equal to or lower than a preset first threshold, is equal to or higher than a second threshold lower than the preset first threshold, and satisfies the first high-temperature pre-start condition. setting the maximum operating time of the air conditioner to the minimum value of the on-time of the air conditioner and a preset maximum on-time of the air conditioner, and activating the air conditioner;
8. The method of claim 7, comprising shutting down the air conditioner in response to an actual on time of the air conditioner being greater than or equal to a maximum operating time of the air conditioner. The first high-temperature preliminary startup condition is
The time when the air conditioner is turned on is reached, the current indoor temperature is greater than a preset third threshold value, and the actual off time of the air conditioner exceeds a preset shortest off time of the air conditioner. 9. The method of claim 8. In the process of controlling the operation of the cooling equipment based on the optimal control parameters,
if the current indoor temperature is greater than the first threshold and the actual off time of the air conditioner is greater than the shortest off time of the air conditioner, set the maximum operating time of the air conditioner to the maximum on time of the air conditioner; and/or activating the air conditioner; and/or deactivating the air conditioner in response to the current indoor temperature being less than the second threshold. The step of controlling operation of the cooling equipment based on the optimal control parameters includes:
activating the heat exchanger in response to meeting a second hot pre-activation condition;
8. The method of claim 7, comprising shutting down the heat exchanger in response to an actual on-time of the heat exchanger being greater than or equal to the on-time of the heat exchanger. When the heat exchanger is an indirect heat exchanger, the second high temperature pre-startup condition is
The heat exchanger is turned on, the current indoor temperature is greater than a preset fourth threshold, and the difference between the current indoor temperature and the outdoor temperature is a preset fifth threshold. or when the heat exchanger is a direct heat exchanger, the second hot pre-startup condition comprises:
A fifth threshold in which the heat exchanger is turned on, the current indoor temperature is greater than a preset fourth threshold, and the difference value between the current indoor temperature and the outdoor temperature is preset. greater than an eighth threshold greater than the value of
The heat exchanger is turned on, the current indoor temperature is greater than a preset fourth threshold, and the difference between the current indoor temperature and the outdoor temperature is a preset eighth threshold. and the current indoor humidity is less than or equal to a preset ninth threshold. the heat exchanger is a direct heat exchanger, and the on-time of the air conditioner and the on-time of the heat exchanger are different;
stopping the heat exchanger in response to turning on the air conditioner;
8. The method of claim 7, further comprising shutting down the air conditioner in response to turning on the heat exchanger. The cooling equipment includes an air conditioner, and after the step of controlling operation of the cooling equipment based on the optimal control parameters,
If the error between the current day's actual operating parameters of the air conditioner and the current day's optimal control parameters of the air conditioner exceeds a tenth threshold value, the current day's optimal control parameters of the air conditioner are determined again. and
4. The method of claim 3, further comprising updating a learning sample data set based on the re-determined optimal control parameters for the current day of the air conditioner. responsive to a currently operating cooling device having a failure and another cooling device being normal, stopping the failed cooling device and turning on the normal cooling device;
2. The method of claim 1, further comprising, responsive to both currently operating two cooling devices failing, activating the failed cooling device when the fault is cleared. Method. In the process of controlling the operation of the cooling equipment based on the optimal control parameters,
The outdoor temperature currently obtained according to the current indoor temperature being lower than a preset sixth threshold and the actual off time of the cooling device being higher than a preset seventh threshold. 2. The method of claim 1, further comprising training the second neural network model based on sample data including room temperature and cooling equipment load. comprising a first processing module, a second processing module, and a control module;
the first processing module is used to determine a current outdoor temperature;
The second processing module is
inputting the synchronized historical sample data of the cooling equipment load and the preset influence factor as a first input parameter into the first neural network model to obtain the today's predicted load of the cooling equipment;
inputting the synchronous history sample data of the outdoor temperature and the predicted load of the cooling equipment on the day as a second input parameter to a second neural network to obtain the predicted indoor temperature on the day;
Inputting the indoor temperature predicted for the day and a preset cooling efficiency factor as a third input parameter into a third neural network to obtain the optimum control parameters for the cooling equipment for the day,
A cooling equipment control device in which the control module is used to control the operation of the cooling equipment based on the optimum control parameters. one or more processors;
a storage device in which one or more programs are stored;
17. Computer equipment for implementing the method of any one of claims 1 to 16 by said one or more processors when said one or more programs are executed by said one or more processors. A computer readable medium on which a computer program is stored,
A computer readable medium on which the method of any one of claims 1 to 16 is implemented by said processor when said computer program is executed by said processor.

Images