Chiller Unit Energy Consumption Calculating Method, Calculating Device, and Chiller Unit

This application claims benefit of Chinese Patent Application No. 202410333704.4, filed Mar. 21, 2024, and all the benefits accruing therefrom under 35 U.S.C. § 119, the contents of which in their entirety are herein incorporated by reference.

This application relates to the field of refrigeration equipment, and specifically relates to a system and method for calculating energy consumption of a chiller unit composed of multiple chillers.

This application provides a system and method for calculating energy consumption of a chiller unit which can predict and calculate the power of a chiller unit with high accuracy and efficiency.

One or more embodiments of this application provides a chiller unit system, the system comprising:

In one or more embodiments, the system further comprises:

In or more embodiments, the processor coupled to the memory storing instructions executable by the processor further causes the controller to:

In one or more embodiments, the cooling load prediction model is a data-driven model, a physical prediction model, or a combination of the data-driven model and the physical prediction model.

In one or more embodiments, the system further comprises:

One or more embodiments of this application further provides a chiller unit system, the system comprising:

In one or more embodiments, wherein the system further comprises:

In one or more embodiments, the processor coupled to the memory storing instructions executable by the processor further causes the controller to:

In one or more embodiments, the cooling load prediction model is a data-driven model, a physical prediction model, or a combination of the data-driven model and the physical prediction model.

In one or more embodiments, the system further comprises:

One or more embodiments of this application further provides a method for calculating energy consumption of a chiller unit, the method comprising:

The technical solutions in the one or more embodiments of this application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this application, and obviously, any single technical feature illustrated or implicit in the drawings of this application still allows any combination or deletion between these technical features (or their equivalents) without any technical obstacles, thereby obtaining other embodiments of this application that my not be directly mentioned herein. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of this application.

Those skilled in the art can appreciate that in order to apply advanced control methods such as model predictive control (MPC) to a centralized heating ventilation air conditioning building system, there are often two forms of using a physical building model and of using a data-driven model to predict refrigeration load or heating load of the centralized heating ventilation air conditioning building system.

The physical building model acquires information such as building type, number of floors, thermal property of the building envelope, climate zone, meteorological parameters, location, orientation, geometry, area, and occupant density, then establishes a physical model and performs energy consumption simulation analysis on the physical model to acquire energy consumption optimization results. However, the physical building model is usually very complex, with many types and requires large calculations, making them unsuitable for control applications.

The data-driven model acquires historical load data, various historical meteorological parameters such as outdoor dry bulb temperature, outdoor relative humidity, outdoor dew point temperature, wind speed, wind direction, cloud cover and atmospheric pressure, as well as weather or chronosystem feature data such as weather forecast, month, day attributes, hours, and on this basis uses algorithms such as multidimensional clustering, multi-step input-multi-step output, attention mechanism to build a training sample database, establishes a data-driven prediction model, trains the model, and then evaluates the trained model, and uses the evaluated model to predict the load of the chiller unit. However, the data-driven model is highly dependent on quality and availability of collected data. In practice, the operating data of different chiller units may have varying qualities and poor data quality, which may lead to inaccurate load prediction of the chiller unit.

Among these, accurate prediction of an input power of the chiller unit is crucial for realizing model predictive control (MPC) of the centralized heating ventilation air conditioning system and improving energy-saving effects.

The chiller unit according to a method for predicting and calculating energy consumption of a chiller unit and a system for predicting and calculating energy consumption of a chiller unit of this application is composed of multiple chillers. A chiller unit composed of three chillers is used as an example for explanation below. The embodiments of this disclosure are not limited to only three chillers, any combination of multiple chillers may be utilized within the scope of the present disclosure.

is a system schematic diagram of a chiller unit according to this application. As shown in, the chiller unit is composed of a first chiller, a second chillerand a third chiller. The first chillerincludes a refrigerant pipeline connected in sequence to a first compressor, a first condenser, a first expansion valveand a first evaporator, forming a refrigerant circulation loop. The high-temperature and high-pressure refrigerant discharged from the first compressorperforms heat exchange with an external medium (such as water or air) in the first condenser, and after being decompressed and expanded by the first expansion valve, performs heat exchange with the external medium (such as water) in the first evaporator, absorbs heat from the external medium (such as water) in the first evaporator, and enters the first compressorto be compressed and heated again, and the cycle continues. The external medium (such as water) in the first evaporatoris cooled after heat being absorbed to form low-temperature cold water of a predetermined temperature, which is driven by the first water pumpand supplied to the user terminal, and returns to the first evaporatorfor circulation again after absorbing heat in the user terminal.

The second chillerincludes a refrigerant pipeline connected in sequence to a second compressor, a second condenser, a second expansion valve, and a second evaporator, forming a refrigerant circulation loop. The external medium (such as water) in the second evaporatoris cooled after heat being absorbed to form low-temperature cold water of a predetermined temperature, which is driven by the second water pumpand supplied to the user terminal, and returns to the second evaporatorfor circulation again after absorbing heat in the user terminal. The third chillerincludes a refrigerant pipeline connected in sequence to a third compressor, a third condenser, a third expansion valve, and a third evaporator, forming a refrigerant circulation loop. The external medium (such as water) in the third evaporatoris cooled after heat being absorbed to form low-temperature cold water of a predetermined temperature, which is driven by the third water pumpand supplied to the user terminal, and returns to the third evaporatorfor circulation again after absorbing heat in the user terminal.

The working principles of the chillerand the chillerare substantially the same as those of the chillerand will not be repeated here.

In addition, the compressor, the compressorand the compressorof the first chiller, the second chillerand the third chilleraccording to this application may be screw type, scroll type or centrifugal type, or any combination thereof. Furthermore, there is no particular limitation on a type of refrigerant used.

The method for predicting and calculating energy consumption of a chiller unit of this application is illustrated below by way of example based on the chiller unit described in.

is a schematic diagram showing steps of a method for predicting and calculating energy consumption of a chiller unit of this application. First, in a chiller unit load prediction step, a total load Q of the chiller unit, i.e., a total load that needs to be provided to a user terminal, is predicted and generated based on a preset data-driven model (i.e., a cooling load prediction model). Then, in a chiller load distribution step, the total load Q of the chiller unit output in the chiller unit load prediction step is distributed to generate a first chiller load Qcorresponding to the first chiller, a second chiller load Qcorresponding to the second chiller, and a third chiller load Qcorresponding to the third chiller, according to a preset chiller load distribution logic.

Subsequently, in a chiller energy efficiency value acquisition step, a first chiller energy efficiency value COP, a second chiller energy efficiency value COPand a third chiller energy efficiency value COPare correspondingly acquired according to a preset chiller load-energy efficiency relationship data, based on the first chiller load Q, the second chiller load Qand the third chiller load Qgenerated by being distributed in the chiller load distribution step.

In a subsequent chiller power calculation step, a first chiller input power P, a second chiller input power Pand a third chiller input power Pare calculated according to the first chiller energy efficiency value COP, the second chiller energy efficiency value COPand the third chiller energy efficiency value COPacquired in the chiller energy efficiency value acquisition step.

is a schematic diagram of a chiller load-energy efficiency curve under specific working conditions, taking the first chilleras an example. As shown in, after the first chiller load Qis generated by distributing in the chiller load distribution step, the first chiller energy efficiency value COPcorresponding to the first chiller load Qcan be acquired by referring to the corresponding working conditions (for example, including an evaporator entering water temperature, a condenser entering water temperature, an outside air temperature, etc.), and the first chiller input power Pcan be calculated accordingly. The working conditions are acquired in real-time via first sensor(s)-operatively connected with a first chiller, second sensor(s)-operatively connected with a second chiller, third sensor(s)-operatively connected with a third chilleror are alternatively based on predetermined data.

Because the chiller load-energy efficiency relationship, i.e., the equipment characteristic curve, is preset with corresponding data when a chiller leaves the factory, the above calculation method can be used to accurately obtain the first chiller input power Pwith the simplest calculated amount.

Similarly, according to the above steps, the second chiller energy efficiency value COPcorresponding to the second chiller load Qcan be calculated, and the second chiller input power Pcan be calculated accordingly. The third chiller energy efficiency value COPcorresponding to the third chiller load Qcan be calculated, and the third chiller input power Pcan be calculated accordingly.

The preset chiller load distribution logic may be an optimal chiller load (OCL) distribution logic, that is, the overall efficiency of the entire chiller unit is maximized and the energy consumption is minimized, by reasonably distributing load to each chiller. For example, a multi-phase genetic algorithm (MPGA), a Lagrangian Algorithm, etc. may be used, and there is no particular limitation. It is sufficient as long as the total load Q of the chiller unit can be distributed into the first chiller load Q, the second chiller load Q, and the third chiller load Q, according to the preset chiller load distribution logic.

Table 1 shows a specific example of the energy efficiency value and input power of each chiller obtained by calculation according to the method for predicting and calculating energy consumption of a chiller unit of this application, taking the total load Q of the chiller unit of 1500 kW as an example.

Further, in a chiller input power summarizing step, the first chiller input power P, the second chiller input power Pand the third chiller input power Pmay be summarized to acquire a total power consumption value P of the chiller unit. For example, as shown in Table 1, 322.23 kW is the total power consumption value P of the chiller unit.

In addition, in some embodiments of this application,is used as an example to illustrate the chiller load-energy efficiency relationship data, and it is not limited to the COP characteristic curve, the energy efficiency ratio EER characteristic curve and the like can also be used as the chiller load-energy efficiency relationship data.

Meanwhile, no matter whether the preset chiller load-energy efficiency relationship data is COP or EER, the COP value and EER value can be calculated using the input power of the entire machine including the compressor of the chiller, pipe valve controller, fan, etc., or using only the power of the compressor alone as the input power, and there is no particular limitation.

Therefore, according to the method for predicting and calculating energy consumption of a chiller unit of this application, by utilizing characteristic curve of each chiller, the accuracy of chiller power prediction can be greatly improved compared to a pure data-driven model.

Because a cooling load of a building as the demand side is only related to the weather and the state of the building itself, and is not related to the specific operating variables (such as chilled water temperature, chiller load rate, etc.) of the chiller as the supply side, predicting the cooling load of the building by using the data-driven model requires fewer variables than predicting chiller energy consumption, and the prediction results are more reliable. Meanwhile, the method for predicting and calculating energy consumption of a chiller unit of this application combines the building cooling load prediction value acquired by using the data-driven model and the accurate chiller characteristic curve data, thereby making prediction of the energy consumption of a chiller more accurate.

is another schematic diagram showing steps of a method for predicting and calculating energy consumption of a chiller unit of this application.

The chiller unit according to some embodiments is the same as that in the other one or more embodiments. The same parts as those in the one or more embodiments are described using the same reference numerals and will not be repeated here.

First, in a water pump load prediction step, a total load q of water pumps is predicted and output based on a preset water pump load prediction model. The total load q of water pumps is related to the total load Q of the chiller unit (i.e., the cooling load of the building) and control variables (such as a setting value of the chilled water temperature and a setting value of supply and return water pressure difference) of the chiller unit that needs to be optimized. Therefore, the water pump load prediction model needs to take the cooling load Q and the control variables of the relevant chiller unit as input to obtain the total load q of the water pumps.

Then, a water pump load distribution step is executed, that is, the total load q of the water pumps is distributed to the first water pumpcorresponding to the first chiller, the second water pumpcorresponding to the second chillerand the third water pumpcorresponding to the third chiller, according to the preset chiller load distribution logic. Further, in a water pump efficiency acquisition step, a first water pump efficiency value η, a second water pump efficiency value ηand a third water pump efficiency value ηare correspondingly acquired according to a preset water pump load-efficiency curve, based on the first water pump load q, the second water pump load qand the third water pump load qdistributed in the water pump load distribution step.

In a water pump power calculation step, the first water pump input power p, the second water pump input power pand the third water pump input power pare calculated according to output results in the water pump load distribution step and the water pump efficiency acquisition step.

is a schematic diagram of a water pump load-efficiency curve, taking the first water pumpas an example. As shown in, after the total water pump load is distributed into the first water pump load q, the second water pump load qand the third water pump load qin the water pump load distribution step, the first water pump efficiency value ηcorresponding to the first water pump load qcan be acquired by referring to the water pump load-efficiency curve corresponding to the first water pump, and the first water pump input power pcan be calculated accordingly.

Because the water pump load-efficiency relationship, as the equipment characteristic curve of the water pump, is preset with corresponding data when a water pump leaves the factory, the above calculation method can be used to accurately obtain the first water pump input power pby a calculated amount.

Similarly, according to the above steps, the second water pump efficiency value ηcorresponding to the second water pump load qcan be easily acquired, and the second water pump input power pcan be calculated accordingly. In addition, the third water pump efficiency value ηcorresponding to the third water pump load qcan be easily acquired, and the third water pump input power pcan be calculated accordingly.

Table 2 shows the flow load distribution and water pump efficiency value of each pump, taking the total load q of the water pumps of 300 kg/s corresponding to the total load Q of the chiller unit as an example. According to the following formula, the corresponding input power of each water pump can be calculated.

water pump input power=flow load×lift×medium density/3600/water pump efficiency