WO2022246627A1

WO2022246627A1 - Method and apparatus for controlling refrigerating device

Info

Publication number: WO2022246627A1
Application number: PCT/CN2021/095676
Authority: WO
Inventors: 卢思亮; 刘宁; 吴宏辉; 谭永宝
Original assignee: 罗伯特·博世有限公司; 卢思亮; 刘宁; 吴宏辉; 谭永宝
Priority date: 2021-05-25
Filing date: 2021-05-25
Publication date: 2022-12-01
Also published as: CN117355710A

Abstract

Provided in the present invention are a method and apparatus for controlling a refrigerating device. The method comprises: training one or more models by using historical data of a refrigerating device, and determining a model from among the one or more models on the basis of a training result; based at least in part on the determined model, controlling a freezing water temperature at an outlet of the refrigerating device; determining whether to perform re-training; and in response to determining to perform re-training, re-training the one or more models by using updated historical data of the refrigerating device, and then determining an updated model from among the one or more models on the basis of a re-training result.

Description

A method and device for controlling refrigeration equipment

technical field

In general, the present invention relates to refrigeration in engineering, and in particular, the present invention relates to methods and apparatus for controlling refrigeration equipment.

Background technique

Today, refrigeration systems are widely deployed in manufacturing plants and buildings around the world, and refrigeration systems usually consume a lot of energy and electricity, which is not conducive to the realization of energy conservation, emission reduction, low carbon emissions, and carbon neutrality. How to make the refrigeration system minimize the energy consumption while meeting the refrigeration demand has been a difficult problem that the engineering community has been facing for a long time.

Thanks to the development of digitalization and sensor technology, many data-driven algorithms and digital building energy systems have been produced, making it easier to monitor, analyze and optimize various energy consumption links in the manufacturing process, so as to be able to show Identify the parts or links that can provide energy savings.

Currently, a class of data-driven optimization methods for cooling systems is devoted to the study of accurate models of energy behavior to control and schedule components in buildings to minimize energy consumption in buildings. For example, the global optimization framework based on the data-driven cooling equipment power prediction model, and the cooling load prediction control through building energy consumption simulation and global optimization, etc. This type of method can minimize the total power consumption of the refrigeration system by using the power prediction model constructed for the main components of the refrigeration system, and under the condition of satisfying the required cooling load and other constraints, the power to be used by each main component To provide a certain degree of energy saving by global optimization, but the quality of its optimization results is usually highly dependent on the prediction model and search space used.

It is desired to provide a more efficient and simple method for optimizing the refrigeration system, so that the energy saving effect can be further improved.

Contents of the invention

A brief overview of one or more embodiments is presented below to provide a basic understanding of these embodiments. This summary is not an extensive overview of all contemplated embodiments, nor is it intended to identify key or critical elements of all embodiments or to delineate the scope of any or all embodiments. Its sole purpose is to present some concepts of one or more embodiments in a simplified form as a prelude to the more detailed description that is presented later.

In one aspect of the present disclosure, there is provided a method for controlling a refrigeration appliance, comprising: using historical data of the refrigeration appliance to train one or more models, and based on the training results from the one or more determining a model in one of the models; controlling outlet chilled water temperature of the cooling unit based at least in part on the determined model; determining whether to perform retraining; and in response to determining that retraining is to be performed, using an update of the cooling unit The one or more models are retrained based on the historical data of the one or more models, and an updated model is determined from the one or more models based on the retraining results.

In another aspect of the present disclosure, there is provided an apparatus for controlling a cooling device, comprising a memory; and at least one processor coupled to the memory and configured to: use the cooling The historical data of the equipment trains one or more models, and determines a model from the one or more models based on the training result; based at least in part on the determined model, controlling the outlet chilled water temperature of the refrigeration equipment; determining whether to retrain; and in response to determining to retrain, retraining the one or more models using updated historical data for the refrigeration appliance, and retraining the one or more models based on retraining results from the one or more The updated model is determined in the model.

In another aspect of the present disclosure, there is provided a computer program product for controlling a refrigeration appliance, comprising processor-executable computer code for: using historical data of the refrigeration appliance to update one or more training a model, and determining a model from the one or more models based on the training results; controlling the outlet chilled water temperature of the refrigeration equipment based at least in part on the determined model; determining whether to perform retraining; and In response to determining that retraining is to be performed, the one or more models are retrained using updated historical data for the refrigeration appliance, and an updated model is determined from the one or more models based on retraining results.

In another aspect of the present disclosure, there is provided a computer readable medium storing computer code for controlling a refrigeration device, the computer code, when executed by a processor, causing the processor to: One or more models are trained using historical data of the refrigeration unit, and a model is determined from the one or more models based on the training results; based at least in part on the determined model, cooling the outlet of the refrigeration unit controlling the water temperature; determining whether to retrain; and in response to determining to retrain, retraining the one or more models using updated historical data for the cooling device, and based on retraining results from the One or more models to determine the updated model.

Other aspects or various modifications of the present disclosure will become more apparent in consideration of the following detailed description and accompanying drawings.

Description of drawings

FIG. 1 illustrates an exemplary block diagram of a refrigeration system 100 according to one or more aspects of the present disclosure;

FIG. 2 is a flowchart illustrating an example method 200 for controlling a refrigeration appliance according to one or more aspects of the present disclosure;

FIG. 3 is a flowchart illustrating an example method 300 for training one or more models according to one or more aspects of the present disclosure;

FIG. 4 is a flowchart illustrating an example method 400 of controlling outlet chilled water temperature of a refrigeration plant based at least in part on a determined model in accordance with one or more aspects of the present disclosure;

FIG. 5 is a flowchart illustrating an example method 500 for determining whether to retrain according to one or more aspects of the present disclosure;

Fig. 6 shows an example of hardware implementation of an apparatus 600 for controlling a cooling device according to one or more aspects of the present disclosure.

Detailed ways

Various embodiments are now described with reference to the drawings, wherein like elements are designated with like reference numerals. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more embodiments. It may be evident, however, that the described embodiments may be practiced without these specific details. In other instances, well-known structures and devices are shown in block diagram form in order to facilitate describing one or more embodiments.

FIG. 1 shows an exemplary block diagram of a refrigeration system 100 . The refrigeration system 100 may include a chiller 102 , a cooling tower 106 in an outer loop 101 , and a room air cooling device 104 in an inner loop 103 . The internal circulation 103 can push the chilled water from the chiller 102 to the interior space of the building for air cooling. Chilled water can absorb heat from the building and return to chiller 102 at a higher temperature. The external circulation 101 can push the condensed water from the refrigerator 102 to the outdoor cooling tower 106 , and the cooling tower 106 releases heat to the outdoor environment, so that the cold water is returned to the refrigerator 102 . The outer loop 101 and the inner loop 103 can run independently of each other. Heat exchange between the inner cycle 103 and the outer cycle 101 may be performed by a refrigerator 102 .

In one aspect of the present disclosure, refrigeration system 100 may also include one or more sensors and a database for collecting and maintaining monitoring data from the one or more sensors. The one or more sensors can be used to monitor one or more parameters for the refrigerator 102, including the chilled water outlet temperature at the water outlet of the refrigerator 102, the return water temperature at the return water outlet of the refrigerator 102, and the refrigeration system The cold water flow and/or velocity in the internal circulation 103 of the 100. In another aspect of the present disclosure, refrigeration system 100 may also include and/or utilize a centralized data platform.

In other aspects of the present disclosure, refrigeration system 100 may also include one or more additional or replacement components.

The present disclosure provides a method for controlling a refrigeration device that uses a data-driven model to control one or more parameters of the refrigeration device, and the model for controlling is not fixed, it can be Changes in response to changes in the external environment (eg, outside temperature) and/or indoor environment (eg, production line layout). Compared to the global optimization-based prior art and methods that typically include building a power prediction model for each major component of a refrigeration system or simulating a building's energy consumption performance, the approach proposed in this disclosure omits the construction of this complex power prediction model Instead, the cooling load prediction model is used, and with the variability of the model, the adaptability to the environment is improved, thereby further improving the energy saving effect of the cooling system.

FIG. 2 is a flowchart illustrating an example method 200 for controlling a refrigeration appliance according to one or more aspects of the present disclosure. Method 200 may be implemented, for example, in refrigeration system 100 shown in FIG. 1 . At step 202, one or more models are trained using historical data of the refrigeration equipment to determine a model from the one or more models. At step 204 , based at least in part on the model determined at step 202 , an outlet chilled water temperature of the refrigeration equipment is controlled. In one aspect of the present disclosure, the control of the cooling device based on the model determined at step 202 is a real-time control and does not require human involvement. At step 206, it is determined whether retraining is to be performed, and in response to the determination that retraining is to be performed, returning to step 202, wherein the one or more models are retrained using updated historical data for the cooling device to learn from the An updated model is determined among the one or more models, or, in response to determining not to perform retraining, return to step 204, wherein the determined model is continued to be used to control the cooling device.

In one aspect of the present disclosure, the refrigeration device in method 200 may include refrigerator 102 as shown in FIG. 1 . In another aspect of the present disclosure, historical data may include one or more time periods of outside temperature and actual cooling loads corresponding to the one or more time periods. For example, the historical data may include the external temperature value recorded hourly in the previous month, and the cooling load actually provided by the refrigeration equipment recorded hourly in the month. In one aspect of the present disclosure, the cooling load actually provided by the refrigeration equipment may be derived based on measurements monitored by one or more sensors deployed in the refrigeration system 100, for example, based on one or more of : the temperature of the chilled water at the water outlet of the refrigerator 102 , the return water temperature at the return water outlet of the refrigerator 102 , and the flow rate and/or flow velocity of the cold water in the internal circulation 103 of the refrigeration system 100 . In one or more aspects of the present disclosure, historical data may be updated monthly, for example, once a month. In other aspects of the present disclosure, historical data may be updated on a period longer or shorter than one month, or on an irregular basis. In another aspect of the present disclosure, the data utilized in method 200 (eg, historical data including actual cooling load and outside temperature, etc.) may be based on a centralized data platform included with and/or coupled to refrigeration system 100 .

FIG. 3 is a flowchart illustrating an example method 300 for training one or more models according to one or more aspects of the present disclosure. For example, method 200 may execute method 300 at step 202 . At step 302, data received from, for example, a database in refrigeration system 100 is pre-processed to generate historical data. In one example, the preprocessing may include downsampling the measured values of one or more sensors collected in the database to obtain downsampled measured values. In another example, preprocessing may include format conversion of the one or more sensor measurements collected in the database. In one aspect of the present disclosure, historical data may be generated by preprocessing to enable the use of measurements from one or more sensors having the same sampling rate and/or format.

At step 304 , one or more models are trained using the historical data generated at step 302 . This historical data can be used as a training and/or testing set for training the model. In one aspect of the present disclosure, the external temperature and the actual cooling load in the previous period may be used as input features of the model, and the cooling load predicted by the model in the next period may be used as the output of the model. In one example, the one or more models may include support vector regression, stochastic regression prediction, decision tree regression, ridge regression, Gaussian process regression, linear regression, Adaboost regression, gradient boosting regression, and the like. In another aspect of the present disclosure, one or more models may be trained and evaluated based on time series cross-validation.

At step 306, based on the training results obtained at step 304, a model is determined from the one or more models. For example, one or more models can be selected with the smallest error. In one example, the metric used to measure error may include mean absolute percent error (MAPE), expressed as:

where y _i is the actual value,

is the predicted result.

In one or more aspects of the present disclosure, method 300 may also be used to determine an updated model. In one example, determining an updated model may include selecting a different model than the current model, for example, updating from a current support vector regression model to a random regression prediction model. In another example, determining an updated model may include updating model parameters without updating the type of model (eg, still using the current support vector regression model).

4 is a flowchart illustrating an example method 400 of controlling outlet chilled water temperature of a chiller based at least in part on a determined model in accordance with one or more aspects of the present disclosure. Method 400 may be performed at step 204 of method 200 . At step 402, the determined model uses the external temperature and the actual cooling load of the previous period as inputs to predict and output the cooling load of the next period. In one aspect of the present disclosure, the model used in step 402 may be determined by method 300 . In another aspect of the present disclosure, the cooling load to be provided by the refrigerator 102 in the next hour can be predicted by using the external temperature of the previous hour and the actual cooling load provided by the refrigerator 102 at the current moment.

At step 404, based on the comparison between the predicted cooling load of the next period and the actual cooling load of the next period, a control instruction for performing fuzzy control on the set point of the outlet chilled water temperature of the refrigeration equipment such as the refrigerator 102 is generated . In one aspect of the present disclosure, it is desirable to maintain a certain amount of cooling load in the refrigeration system 100 to meet the cooling demands within the manufacturing plant and/or building. However, as factors such as external temperature change, the amount of power consumed to provide the same cooling load may vary, providing opportunities for power savings and energy savings. For example, when the cooling load for the next period predicted by the determined model based on the external temperature and the actual cooling load for the previous period is greater than the actual cooling load for the next period, a further next period for increasing the refrigerator 102 is generated The control command of the set point of the outlet chilled water temperature to save unnecessary power consumption; or, when the cooling load for the next period predicted by the determined model based on the external temperature and the actual cooling load in the previous period is less than the next period When the actual cooling load is in a period, a control instruction for reducing the set point of the outlet chilled water temperature of the refrigerator 102 in the next period is generated to ensure that the cooling demand can be met.

At step 406 , based on the control instruction generated at step 404 , the set point of the chilled water temperature at the outlet of the refrigerator 102 in the next period is adjusted. For example, the outlet chilled water temperature of the refrigerator 102 in the next period can be adjusted by adjusting the programmable logic controller (PLC) used for the outlet chilled water temperature set point of the refrigerator 102 . In one example, the temperature of the chilled water at the outlet of the refrigerator 102 for the next hour may be increased or decreased in steps of 0.25°C (Celsius) according to the control instruction. In one aspect of the present disclosure, the temperature of the chilled water at the outlet of the refrigerator 102 can be initially set at a predetermined set point, and the water temperature can be adjusted within a certain range (for example, 8°C to 12°C ).

At step 408, the actual cooling load for each period is calculated using the measured values monitored by one or more sensors deployed in the refrigeration system 100, and the calculated actual cooling load is used as the next Input to step 402 and step 404 in the loop.

In one aspect of the present disclosure, steps 402 to 408 may be repeated according to a predetermined cycle (eg, every hour), so as to realize real-time control of the refrigerator 102 . In other aspects of the present disclosure, steps 402 to 408 may also be repeated in an aperiodic manner.

FIG. 5 is a flowchart illustrating an example method 500 for determining whether to retrain according to one or more aspects of the present disclosure. Method 500 may be performed at step 206 of method 200 . At step 502, an error is calculated between the cooling load predicted by the determined model for one or more time periods and the actual cooling load corresponding to the one or more time periods. In one example, the one or more predicted cooling loads predicted at step 402 of method 400 for one or more time periods (eg, hourly) are compared with the one or more predicted cooling loads at step 408 of method 400 for the one or more One or more cooling load actual values measured by sensors for a period of time are compared to calculate the error between the model prediction value and the actual measured value. This error can be expressed by mean absolute percent error (MAPE). In one example, the error may be calculated periodically, for example, once a month. In another example, the error may also be calculated aperiodically.

At step 504, the calculated error is compared with an error threshold (eg, 5%) to determine whether to perform retraining. For example, when the calculated error is greater than the error threshold, it is determined to perform retraining, and when the calculated error is less than or equal to the error threshold, it is determined not to perform retraining so as to continue using the model determined by the method 300 .

Suppose, using the method 400 shown in FIG. 4 , the actual cooling load for period k-1 derived via step 408 is 1 refrigeration ton (RT) and the actual cooling load for period k is 0.8 refrigeration ton (RT), and Used as input for

steps

402 and 404 in the next cycle (ie, for k+1 epochs). At step 402 , based on the actual cooling load of 1 RT during the k-1 period and the external temperature during the k-1 period, the determined model is used to predict that the cooling load during the k period is 1.1 RT. At step 404, since the predicted 1.1 refrigerated tons (RT) for the k period is greater than the actual cooling load of 0.8 refrigerated tons (RT) for the k period, a control instruction for increasing the outlet chilled water temperature is generated. At step 406, according to the control instruction, the temperature of the outlet chilled water in the k+1 period is increased by a step size of 0.25°C. At step 408, the actual cooling load of the period k+1 obtained by sensor measurement is 0.7 RT, and 0.8 RT and 0.7 RT are used for the next cycle (i.e., For the input of step 402 and step 404 in k+2 period). At step 402 , based on 0.8 refrigerating tons (RT) and the external temperature of k period, the cooling load of k+1 period is predicted to be 1.2 refrigerating tons (RT) by using the determined model. At step 404, since the predicted 1.2 refrigerating tons (RT) for the k+1 period is greater than the 0.7 refrigerating tons (RT) of the actual cooling load for the k+1 period, a control instruction for increasing the outlet chilled water temperature is generated. At step 406, according to the control instruction, the temperature of the outlet chilled water in the period k+2 is increased by a step size of 0.25°C. At step 408 , the actual cooling load for the period k+2 obtained by sensor measurement is 0.6 tons of refrigeration (RT). Returning to step 402, based on the actual cooling load of 0.7 refrigeration tons (RT) in the k+1 period and the external temperature in the k+1 period, the determined model is used to predict that the cooling load in the k+2 period is 1.3 refrigeration tons ( RT). The above example can be represented by Table 1.

Table 1

the

k-1 period

k period

k+1 period

k+2 period

冷负荷的预测值Predicted value of cooling load	the	1.11.1	1.21.2	1.31.3
冷负荷的实际值Actual value of cooling load	11	0.80.8	0.70.7	0.60.6

It can be seen from Table 1 that the relative error percentage between the actual value (0.8) and the predicted value (1.1) of period k is about 38% (that is, (1.1-0.8)/0.8), and the actual value of period k+1 ( The relative percent error between 0.7) and the predicted value (1.2) was 71% (ie, (1.2-0.7)/0.7). The error for both time periods has exceeded an error threshold (eg, 5%). If the average relative error percentage exceeds the error threshold over a certain period of time (e.g., one month), this may mean that the currently used model has a large error in the prediction results for changes in factors such as external temperature, or is no longer applicable in the current changing environment. For example, in the example in Table 1, possible situations include that the external temperature rises rapidly, so that the chilled water temperature required to provide the required cooling load needs to continue to decrease. However, due to the inadaptability of the currently used model, the outlet frozen water temperature in the k+1 period and k+2 period is even increased. Therefore, as described in the method 500 in FIG. 5 , when the calculated average error MAPE is greater than the error threshold, it is determined to perform retraining to determine a more suitable model.

In one aspect of the present disclosure, according to the example in Table 1, method 300 may use updated historical data, for example, including actual cooling loads 1, 0.8, 0.7, 0.6RT, and the external temperature corresponding to these time periods, to retrain one or more models, so that the updated model is compared with the currently used model (the error between the predicted value and the actual value is larger, for example, 38 % and 71%) are better able to adapt to recent environmental changes.

According to one or more aspects of the present disclosure, by performing one or more of method 200, method 300, method 400, and/or method 500, historical data and external temperature for the refrigeration equipment itself may be used to provide The prediction of cooling load, and through the retraining of the model, enables the model to adapt to the changing environment.

Fig. 6 shows an example of hardware implementation of an apparatus 600 for controlling a cooling device according to one or more aspects of the present disclosure. The apparatus 600 for controlling a refrigeration device may include a memory 610 and at least one processor 620 . The processor 620 may be coupled to the memory 610 and configured to perform one or more of the method 200, the method 300, the method 400 and/or the method 500 described above with reference to FIGS. 2, 3, 4 and 5 . The processor 620 may be a general-purpose processor, or may also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors combined with a DSP core, or other such structures. Memory 610 may store input data, output data, data generated by processor 620 and/or instructions executed by processor 620 .

The various operations, models and networks described in connection with this disclosure may be implemented as hardware, software executed by a processor, firmware, or any combination thereof. According to one or more aspects of the present disclosure, a computer program product for controlling a refrigeration device may include a method for performing the method 200, the method 300, the method 400 and the method described above with reference to FIGS. A processor of one or more of methods 500 may execute computer code and/or. According to other aspects of the present disclosure, a computer-readable medium may store computer codes for controlling a cooling device, which, when executed by a processor, may cause the processor to perform the operations described above with reference to FIGS. 2 , 3 , 4 and 5 . One or more of method 200, method 300, method 400, and/or method 500. Computer-readable media includes both non-transitory computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. Any connection is properly termed a computer-readable medium.

The above description of the present disclosure is provided to enable any person skilled in the art to use or practice the various embodiments. Various modifications to the above-described embodiments will be readily apparent to those skilled in the art, and the rationale defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the scope of the claims is not intended to be limited to the embodiments disclosed herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

A method for controlling refrigeration equipment comprising:

training one or more models using historical data of the refrigeration appliance, and determining a model from the one or more models based on training results;

controlling an outlet chilled water temperature of the refrigeration unit based at least in part on the determined model;

determine whether to retrain; and

In response to determining that retraining is to be performed, the one or more models are retrained using updated historical data for the refrigeration appliance, and an updated model is determined from the one or more models based on retraining results.
The method of claim 1, wherein the historical data includes outside temperature for one or more time periods and actual cooling loads corresponding to the one or more time periods, and wherein the updated historical data includes The external temperature for one or more time periods after the one or more time periods and the actual cooling load corresponding to the one or more time periods after the one or more time periods.
The method of claim 1, wherein said controlling further comprises:

Based on the comparison between the cooling load of the next period predicted by the determined model and the actual cooling load of the next period, fuzzy control is performed on the set point of the outlet chilled water temperature of the refrigeration equipment.
The method of claim 1, wherein:

Whether to perform retraining is determined based on an error between a cooling load predicted by the determined model for one or more time periods and an actual cooling load corresponding to the one or more time periods.
The method according to claim 3, wherein,

When the cooling load of the next period predicted by the determined model is greater than the actual cooling load of the next period, increasing the set point of the outlet chilled water temperature of the refrigeration equipment for the next period; or

When the cooling load of the next period predicted by the determined model is smaller than the actual cooling load of the next period, the set point of the chilled water temperature at the outlet of the refrigeration equipment in the next period is lowered.
The method of claim 4, wherein the error comprises the mean absolute difference between the cooling load predicted by the determined model for one or more time periods and the actual cooling load corresponding to the one or more time periods. Percent Error (MAPE); and where,

When the MAPE is greater than a threshold, it is determined that retraining is to be performed, and when the MAPE is less than or equal to a threshold, it is determined not to perform retraining, so as to continue using the determined model.
The method of claim 1, wherein the one or more models comprise one or more of the following:

Support Vector Regression, Stochastic Regression Forecasting, Decision Tree Regression, Ridge Regression, Gaussian Process Regression, Linear Regression, Adaboost Regression, Gradient Boosting Regression.
An apparatus for controlling refrigeration equipment comprising:

storage; and

At least one processor coupled to the memory and configured to perform the method of any one of claims 1-7.
A computer program product for controlling a refrigeration appliance comprising processor-executable computer code for performing the method of any one of claims 1 to 7.
A computer-readable medium storing computer code for controlling a refrigeration device, the computer code, when executed by a processor, causing the processor to perform the method of any one of claims 1-7.