Industrial park energy efficiency monitoring method and system based on digital twinning

Abstract

The present application relates to the field of energy efficiency monitoring, and more particularly to an industrial park energy efficiency monitoring method and system based on digital twinning, the method comprising: collecting source heat shock temperature, cooling water flow, pipe network environment temperature and actual power consumption of the water chiller, then calculating the reference temperature through the sliding time window, combining the current temperature difference and the cooling water flow to obtain the transient heat power deviation degree to quantify the heat load disturbance, then combining the pipe network parameters and the cooling water flow to calculate the heat flow transmission delay step and obtain the dynamic heat shock intensity, finally calculating the drift correction factor through the intensity, adjusting the ARIMA model coefficient, predicting the expected energy consumption baseline and comparing the actual power consumption to judge the unit operation state. The present application realizes accurate energy efficiency monitoring by sensing the upstream working condition, restoring the pipe network transmission characteristics and adaptively correcting the ARIMA model, which fundamentally avoids false alarms.

Description

Technical Field

[0001] This invention relates to the field of energy efficiency monitoring, and more particularly to a method and system for monitoring energy efficiency in industrial parks based on digital twins. Background Technology

[0002] Digital twins integrate digital modeling, real-time data acquisition, and dynamic simulation analysis, enabling precise mapping, monitoring, and projection of the entire lifecycle state of physical entities. By breaking down information barriers between digital and physical entities, digital twins are a core technology for industrial digital transformation and an inevitable choice for achieving energy conservation, emission reduction, and improving the level of intelligent energy management.

[0003] In the actual operation of the cooling water system in the industrial park, the cooling water, as the direct carrier and moving body of the heat load disturbance, flows and is transmitted in the cooling pipe network and return water pipe, and there is a delay in heat flow transmission. Ultimately, it acts on the chiller unit and directly determines its operating load and energy consumption status.

[0004] Existing energy efficiency monitoring technologies in industrial parks largely rely on statistical algorithms such as ARIMA. These algorithms predict the expected energy consumption baseline by fitting historical energy consumption time-series data from end-point equipment, and then determine energy efficiency anomalies through residual comparison. This approach has significant application flaws. Because the ARIMA algorithm heavily relies on the assumption of time series stationarity and lacks situational awareness of upstream production processes, it cannot perceive or quantify the delayed physical impact of sudden changes in production scheduling within the industrial park on upstream heat loads. Furthermore, it fails to consider the time delay and flow rate fluctuations in heat transfer within the cooling pipe network, resulting in a significant lag between the predicted expected energy consumption baseline and actual equipment operating conditions. When a sudden heat load occurs upstream, actual energy consumption changes abruptly, while the predicted baseline remains stable. This drastically amplifies the residual difference, leading to frequent false energy efficiency anomaly alarms and severely disrupting normal equipment operation and energy scheduling within the park. Summary of the Invention

[0005] To address the problem that existing ARIMA algorithms for energy efficiency monitoring in industrial parks rely on time-series stationarity, lack upstream sensing capabilities, and fail to consider the heat flow characteristics of cooling pipe networks, resulting in delayed predictions of expected energy consumption baselines, which can easily trigger false alarms and interfere with park operation and maintenance scheduling, this invention provides solutions in the following aspects.

[0006] In the first aspect, the energy efficiency monitoring method for industrial parks based on digital twins includes: acquiring the source thermal shock temperature, source cooling water flow rate, ambient temperature of the cooling network of the return water main, and actual power consumption at the input of the chiller unit's distribution cabinet at the water collection point of the production workshop; setting a sliding time window, calculating the average of the historical return water temperature sequence within the sliding time window as a reference temperature, and calculating the degree of transient heat power deviation based on the difference between the current return water temperature and the reference temperature and the current cooling water flow rate, thus quantifying the heat load disturbance caused by sudden changes in upstream production scheduling; combining the cooling network parameters and the current cooling water flow rate, calculating the heat flow transmission delay step and the actual spatial velocity, and based on the heat flow transmission delay step... The dynamic thermal shock intensity at the current moment is calculated by considering the transient thermal power deviation, heat flow delay step, and normalized actual spatial velocity at the corresponding moment of the long-matched model. Based on the dynamic thermal shock intensity, the drift correction factor at the current moment is calculated. The first-order autoregressive coefficients of the ARIMA model are adjusted using the drift correction factor. The adjusted first-order autoregressive coefficients are then substituted into the ARIMA model to obtain the corrected prediction model. This model predicts the expected energy consumption baseline of the chiller unit at the next moment, calculates the residual between the actual power consumption of the chiller unit and the expected energy consumption baseline, and determines the operating status of the chiller unit. The product of the first-order autoregressive coefficients of the traditional ARIMA model and 1 plus the drift correction factor is used as the adjusted first-order autoregressive coefficients.

[0007] Preferably, the calculation method for the degree of deviation of the transient thermal power is as follows:

[0008] Calculate the difference between the current return water temperature and the reference temperature, multiply the difference by the current cooling water flow rate, and then calculate the degree of transient thermal power deviation using the hyperbolic tangent function.

[0009] Preferably, the heat flow transfer delay step is calculated as follows:

[0010] The cooling network parameters include the total length and cross-sectional area of ​​the pipes from the front-end thermal node to the inlet of the terminal chiller unit. The ratio of the current cooling water flow rate to the pipe cross-sectional area is used as the actual spatial velocity. The ratio of the total pipe length to the actual spatial velocity is rounded down to obtain the heat flow transmission delay step.

[0011] Preferably, the dynamic thermal shock intensity is calculated as follows:

[0012] The actual spatial velocity at the current moment is normalized, and the product of the normalized actual spatial velocity and the degree of deviation of the transient thermal power is taken as the dynamic thermal shock intensity at the current moment.

[0013] Preferably, the drift correction factor is calculated as follows:

[0014] The difference between the current dynamic thermal shock intensity and the mean dynamic thermal shock intensity within the corresponding sliding time window is taken as the dynamic thermal shock deviation; the average absolute value of the dynamic thermal shock intensity at each historical moment within the sliding time window is taken as the historical average thermal shock intensity, and the sum of the historical average thermal shock intensity and the minimum hyperparameter is taken as the denominator; the dynamic thermal shock deviation is divided by the denominator to obtain the drift correction factor at the current moment.

[0015] Preferably, the steps for determining the operating status of the chiller unit include:

[0016] The ARIMA prediction model is updated based on the adjusted first-order autoregressive coefficients to obtain the expected energy consumption baseline of the chiller unit at the current moment. The difference between the actual power consumption of the chiller unit at the current moment and the expected energy consumption baseline is taken as the residual at the current moment. The standard deviation of the historical residual sequence formed by the residuals of the chiller units at each historical moment within the corresponding sliding time window at the current moment is calculated.

[0017] If the residual at the current moment is greater than 3 times the standard deviation, the system determines that the chiller unit has significantly high energy consumption behavior and outputs an inefficiency alarm. Conversely, if the residual at the current moment is less than or equal to 3 times the standard deviation, the system determines that the current energy consumption fluctuation is within the normal range and does not trigger an alarm.

[0018] Secondly, an industrial park energy efficiency monitoring system based on digital twins includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned industrial park energy efficiency monitoring method based on digital twins is implemented.

[0019] The present invention has the following effects:

[0020] 1. This invention quantifies the heat load disturbance caused by sudden changes in upstream production scheduling in real time by measuring return water temperature and cooling water flow rate. It breaks through the dependence of traditional ARIMA models on the stationarity of time series, realizes accurate perception of changes in production conditions, and solves the problem of the disconnect between energy consumption prediction and actual operating conditions from the source.

[0021] 2. This invention combines the physical parameters of the cooling pipe network with the fluid flow velocity to calculate the heat transfer delay step, restores the heat transfer characteristics in the pipe network, and enables the dynamic thermal shock intensity to truly reflect the hysteretic impact of upstream disturbances on the chiller unit, significantly improving the real-time performance and accuracy of the expected energy consumption baseline prediction.

[0022] 3. This invention calculates the drift correction factor by dynamically adjusting the first-order autoregressive coefficient of the ARIMA model through dynamic thermal shock intensity calculation, so that the expected energy consumption baseline can dynamically rise and fall with positive or negative thermal load shocks, effectively distinguishing between normal load fluctuations and actual energy efficiency anomalies, fundamentally avoiding false alarms, and ensuring the stable and reliable operation and maintenance of park equipment and energy dispatch. Attached Figure Description

[0023] Figure 1 This is a flowchart of steps S1-S4 in the digital twin-based industrial park energy efficiency monitoring method of this invention.

[0024] Figure 2 This is a structural block diagram of an industrial park energy efficiency monitoring system based on digital twins, according to an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0026] Specific implementation scenario: In the actual operation of the cooling system in the industrial park, the production workshop serves as the heat source. The high-temperature waste heat generated by the machining operation of the internal equipment is absorbed by the cooling water and collected in the return water main of the production workshop, forming a high-temperature cooling water flow carrying information on heat load changes. Driven by the variable frequency water pump of the cooling pipe network, the high-temperature cooling water flow flows along the cooling pipe network and the conveying pipeline. After a certain transmission path, it reaches the terminal chiller unit, where the evaporator of the chiller unit performs heat exchange and cooling, ultimately maintaining the thermal balance of the system.

[0027] In this process, cooling water, as the direct carrier and moving body of heat load disturbance, is transmitted in the cooling pipe network and return water pipe and forms a heat flow transmission delay, which ultimately acts on the chiller unit and determines its energy consumption and operating status. Traditional methods only monitor the end energy consumption and ignore this whole-link physical process, which can easily cause prediction errors and false alarms.

[0028] Reference Figure 1 The energy efficiency monitoring method for industrial parks based on digital twins includes steps S1-S4, as detailed below:

[0029] S1: Obtain the source thermal shock temperature of the return water main collection point in the production workshop, the source cooling water flow rate, the ambient temperature of the cooling pipe network of the return water main, and the actual power consumption at the input terminal of the chiller unit's power distribution cabinet.

[0030] Temperature sensors are installed at the junctions where the return water pipes converge into the main pipe in the production workshop. These sensors collect the real-time temperature of the cooling water at these junctions, serving as the source thermal shock temperature and directly reflecting changes in heat load caused by sudden changes in production schedules of upstream equipment. Flow sensors are installed on the return water main at the same junction, synchronously collecting the instantaneous flow rate of the cooling water at the junction, which is the source cooling water flow rate, used to quantify heat load transfer efficiency. Ambient temperature sensors are attached to the outer wall of the return water main to collect the ambient temperature around the pipe, serving as the ambient temperature of the cooling network in the return water main, used to help correct for temperature loss errors during heat transfer. Power meters are installed at the power input of the chiller unit's distribution cabinet to collect the actual power consumption data during chiller unit operation, which is the actual power consumption at the chiller unit's distribution cabinet input, serving as core data for judging the unit's operating status. All collected data is transmitted in real-time to the digital twin platform via a data acquisition module, enabling comprehensive monitoring of the entire energy efficiency testing process in the industrial park.

[0031] S2: Set a sliding time window, calculate the average value of the historical return water temperature sequence within the sliding time window as the benchmark reference temperature, and calculate the degree of transient thermal power deviation based on the difference between the current return water temperature and the benchmark reference temperature and the current cooling water flow rate, thereby quantifying the heat load disturbance caused by sudden changes in upstream production scheduling.

[0032] The difference between the current return water temperature and the reference temperature is calculated. The difference reflects the degree of deviation of the current heat load from the recent average operating conditions. A positive difference indicates that the heat load is higher than the average level, and a negative difference indicates that the heat load is lower than the average level. In addition, the magnitude of heat power is not only related to temperature changes, but also closely related to the cooling water flow rate. Under the same temperature deviation, the larger the flow rate, the higher the heat load transmission efficiency and the stronger the impact on the chiller unit. The difference is multiplied by the current cooling water flow rate to obtain the preliminary quantified value of heat power deviation. In order to avoid the quantified value being abnormally large due to extreme temperature deviation or flow fluctuation, the transient heat power deviation is calculated by the hyperbolic tangent function.

[0033] In this embodiment, the sliding time window The sliding time window is divided backward from the last data point of the current moment to form the historical backwater temperature sequence for the current moment. Specifically, the sliding time window can be adjusted according to the specific circumstances.

[0034] Specifically, the degree of deviation of transient thermal power satisfies the following relationship:

[0035] ;

[0036] In the formula, Indicates the current time The degree of deviation of transient thermal power; Indicates the current time The return water temperature; This indicates the current reference temperature. Indicates the current time Cooling water flow rate This represents the hyperbolic tangent function.

[0037] In other words, firstly, changes in heat load caused by sudden changes in production scheduling will be immediately reflected in the return water temperature and flow rate at the corresponding locations, indicating a close physical relationship. Secondly, sudden changes in heat load can be broken down into changes in waste heat release and waste heat transfer efficiency. This indicates the degree of temperature deviation between the current upstream production status and the recent normal thermal balance. It indicates the ability of cooling water to transport the current upstream production status, breaking the recent normal thermal balance state. It quantifies the degree of transient thermal power deviation by comprehensively considering both the magnitude of thermal load disturbance and the transmission efficiency.

[0038] If the current return water temperature is greater than or equal to the current reference temperature, the industrial park's production workshop may be under a sudden high-intensity processing task, with equipment operating at full load. At this time, the high-temperature waste heat discharged into the cooling pipe network increases sharply, the transient heat power deviation is positive, and the greater the transient heat power deviation, the greater the transient heat power deviation.

[0039] If the return water temperature at the current moment is lower than the reference temperature at the current moment, the production workshop in the industrial park may be in the process of batch handover, with a large number of equipment switching to standby mode or the production line suddenly shutting down due to a malfunction. At this time, the high-temperature waste heat discharged into the cooling pipe network drops sharply, and the system has cold capacity redundancy. At this time, the transient thermal power deviation is negative, and the transient thermal power deviation decreases as the temperature deviation increases. The larger the cooling water flow rate, the smaller the transient thermal power deviation.

[0040] S3: Combining the cooling pipe network parameters and the current cooling water flow rate, calculate the heat transfer delay step and the actual spatial velocity. Based on the corresponding transient thermal power deviation at the time of the heat transfer delay step matching, the heat transfer delay step, and the normalized actual spatial velocity, calculate the dynamic thermal shock intensity at the current time.

[0041] When heat flows through the cooling pipe network, there is a certain time delay due to the length of the pipes. This delay directly affects the timing when upstream heat load disturbances act on the chiller unit. If this delay is ignored, the calculation of dynamic thermal shock intensity will become disconnected from the actual operating conditions, thus affecting the accuracy of subsequent model corrections. Therefore, it is necessary to calculate the heat flow transmission delay step size first. The specific steps are as follows:

[0042] Obtain the total length of the pipeline from the front-end thermal node (the water collection point of the return water main in the production workshop) to the inlet of the terminal chiller unit (unit: ) and the cross-sectional area of ​​the pipe (unit) The cross-sectional area of ​​the pipe is predetermined based on the pipe specifications. Since the actual spatial velocity of the cooling water is directly related to the flow rate and the cross-sectional area of ​​the pipe, the larger the flow rate and the smaller the cross-sectional area of ​​the pipe, the faster the flow rate. Therefore, the ratio of the current cooling water flow rate to the cross-sectional area of ​​the pipe is used as the actual spatial velocity, which can truly reflect the speed of heat flow transmission. The ratio of the total pipe length to the actual spatial velocity is used as the theoretical time required for heat flow to be transmitted from the front thermal node to the chiller inlet. The theoretical time is rounded down to obtain the heat flow transmission delay step.

[0043] The actual spatial velocity at the current moment is normalized, and the product of the normalized actual spatial velocity and the degree of deviation of the transient thermal power is taken as the dynamic thermal shock intensity at the current moment.

[0044] Specifically, the dynamic thermal shock intensity satisfies the following relationship:

[0045] ;

[0046] In the formula, Indicates the current time Dynamic thermal shock intensity; Indicates the first The degree of deviation of transient thermal power at any given moment; Indicates the current time The heat flow transport delay step; Indicates the current time The actual spatial velocity of the fluid medium; This represents the standard normalization function.

[0047] According to the principles of fluid mechanics, if the transient thermal power deviation is positive at the current moment and the actual spatial velocity of the fluid medium is relatively high, the high-temperature heat flow will rush into the unit at a high velocity under the strong pressure of the variable frequency water pump in the cooling pipe network, instantly forming a large temperature gradient difference on the inner wall of the evaporator. Conversely, if the actual spatial velocity of the fluid medium is slow at the current moment, the chiller unit has sufficient time to smoothly adjust the refrigeration system, the unit's operating conditions change gently, and the system operates stably.

[0048] S4: Calculate the drift correction factor at the current moment based on the dynamic thermal shock intensity, adjust the first-order autoregressive coefficient of the ARIMA model through the drift correction factor, substitute the adjusted first-order autoregressive coefficient into the ARIMA model to obtain the corrected prediction model, predict the expected energy consumption baseline of the chiller unit at the next moment, calculate the residual between the actual power consumption of the chiller unit and the expected energy consumption baseline, and determine the operating status of the chiller unit.

[0049] The core function of the drift correction factor is to quantify the deviation of the current dynamic thermal shock intensity from the recent average level, thereby providing a basis for the adaptive adjustment of the first-order autoregressive coefficients of the ARIMA (Auto Regressive Integrated Moving Average) model. This solves the problem that the traditional ARIMA model cannot adapt to dynamic fluctuations in heat load and has prediction lag. The specific steps are as follows:

[0050] Based on the previously set sliding time window, the dynamic thermal shock intensity at the current moment is used as the benchmark, and the difference between the current dynamic thermal shock intensity and the mean dynamic thermal shock intensity within the corresponding sliding time window is used as the dynamic thermal shock deviation. The average absolute value of the dynamic thermal shock intensity at each historical moment within the sliding time window is used as the historical average thermal shock intensity. To prevent the historical average thermal shock intensity from approaching 0 and causing calculation failure, a minimum hyperparameter is introduced, and the sum of the historical average thermal shock intensity and the minimum hyperparameter is used as the denominator. The dynamic thermal shock deviation is divided by the denominator to obtain the drift correction factor at the current moment, which reflects the degree and direction of the deviation of the current thermal shock from the recent average level. This provides an accurate basis for the subsequent positive amplification or negative weakening adjustment of the first-order autoregressive coefficients of the ARIMA model, realizing the adaptive adaptation of the model to dynamic fluctuations in heat load.

[0051] Specifically, the drift correction factor satisfies the following relationship:

[0052] ;

[0053] In the formula, Indicates the current time The drift correction factor; Indicates the current time Dynamic thermal shock intensity; Indicates the current time The average dynamic thermal shock intensity within the corresponding sliding time window; Indicates the length of the sliding time window; Indicates the time sequence number within the sliding time window; Indicates the current time The corresponding sliding time window Dynamic thermal shock intensity at a historical moment; Indicates the absolute value symbol; For example, the hyperparameter for the minimum value To avoid the denominator being 0.

[0054] When the upstream workshop suddenly experiences high-intensity production and the equipment is running at full load, the return water temperature rises significantly. The cooling water carrying high-temperature waste heat forms a positive thermal shock and acts on the chiller unit.

[0055] Dynamic thermal shock intensity at this moment Greater than the average level within the sliding time window This causes dynamic thermal shock deviation If the value is positive, the drift correction factor is calculated assuming the denominator is always positive. Outputting a positive value enables positive correction of the ARIMA model, allowing the expected energy consumption baseline to quickly follow the increase in unit load.

[0056] When a large number of upstream equipment are put into standby or shutdown mode, the return water temperature drops sharply, and the cooling water has excess cooling capacity, resulting in a negative cooling impact.

[0057] Dynamic thermal shock intensity at this moment Less than the average level within the sliding time window This causes dynamic thermal shock deviation Even if the value is negative, the denominator remains positive, therefore the drift correction factor... Outputting negative values ​​enables negative correction of the ARIMA model, ensuring that the expected energy consumption baseline follows the unit load decrease in a timely manner.

[0058] In other words, by using the positive or negative sign of the dynamic thermal shock deviation, the drift correction factor can adaptively identify whether the upstream heat load is a positive surge or a negative drop, and thereby drive the ARIMA model to achieve dynamic correction consistent with the actual shock direction, so that the expected energy consumption baseline always fits the actual operating conditions of the chiller unit.

[0059] The product of the first-order autoregressive coefficients of the traditional ARIMA model and 1 plus the drift correction factor is used as the adjusted first-order autoregressive coefficients.

[0060] Specifically, the formula for adjusting the first-order autoregressive coefficients satisfies the following relationship:

[0061] ;

[0062] In the formula, This represents the adjusted first-order autoregressive coefficient; This represents the first-order autoregressive coefficients obtained by fitting based on the stationarity assumption using traditional methods; Indicates the current time The drift correction factor.

[0063] The product of the first-order autoregressive coefficient of the traditional ARIMA model and 1 plus the drift correction factor is used as the adjusted first-order autoregressive coefficient. Its physical and logical significance is that it enables the model to adaptively adjust the influence weight of historical energy consumption on the current prediction according to the direction and intensity of upstream heat load impact, thereby breaking through the limitation of the traditional ARIMA stationarity assumption and realizing the dynamic tracking of the expected energy consumption baseline.

[0064] When the system is subjected to a positive thermal shock When the drift correction factor is positive, it will amplify the adjusted first-order autoregressive coefficient, making the model more dependent on the upward trend of energy consumption in the previous moment, and quickly raising the energy consumption prediction baseline at the current moment. This will make the predicted value match the actual high energy consumption state of the chiller unit caused by the high temperature heat flow impact, avoid misjudging the equipment as inefficient operation due to the low prediction, and thus eliminate false abnormal alarms.

[0065] When the system is subjected to a negative thermal shock A negative drift correction factor reduces the adjusted first-order autoregressive coefficient, thereby reducing the model's dependence on the previous high-energy-consumption state. This drives the energy consumption prediction baseline to decrease steadily with the cooling capacity redundancy, aligning with the actual load reduction of the chiller unit. This avoids misjudging equipment failure or downtime due to overestimation, thus eliminating false anomaly warnings.

[0066] Under stable operating conditions The drift correction factor is approximately 0, and the adjusted first-order autoregressive coefficient is basically consistent with the traditional coefficient. The model maintains the state of conventional time series prediction, ensuring the prediction stability under normal working conditions.

[0067] It should be noted that the first-order autoregressive term in the original model is replaced with the adjusted first-order autoregressive coefficient, while the higher-order autoregressive terms remain unchanged.

[0068] The ARIMA prediction model is updated based on the adjusted first-order autoregressive coefficients, enabling the model to adapt to the dynamic fluctuations of the current heat load and no longer be limited by the traditional stationarity assumption. This outputs the expected energy consumption baseline of the chiller unit at the current moment. The baseline is the theoretical energy consumption value obtained based on the current thermal shock condition, historical energy consumption data, and adaptive model adjustments, accurately reflecting the normal energy consumption level of the chiller unit under the current conditions. Using any given moment as the target moment, the difference between the actual power consumption of the chiller unit at the target moment (obtained through equipment monitoring) and the expected energy consumption baseline is taken as the residual at the target moment. The magnitude of the residual directly reflects the degree of deviation between the actual energy consumption and the expected energy consumption baseline; the larger the residual, the more significant the deviation between the actual energy consumption and the theoretical normal energy consumption. The standard deviation of the historical residual sequence within the corresponding sliding time window at the current moment is calculated as a reference standard for judging whether the current residual belongs to an abnormal deviation. The unit's operating status is determined based on the 3σ principle (a commonly used anomaly judgment standard in industrial monitoring).

[0069] If the residual at the current moment is greater than 3 times the standard deviation, it indicates that the current actual energy consumption deviates significantly from the normal baseline. In this case, the system determines that the chiller unit has significantly high energy consumption behavior and outputs an equipment inefficiency alarm to remind maintenance personnel to promptly investigate equipment faults and optimize operating parameters. Conversely, if the residual at the current moment is less than or equal to 3 times the standard deviation, it indicates that the current energy consumption fluctuation is within the recent normal fluctuation range and is a normal load fluctuation caused by sudden changes in upstream production scheduling, rather than an abnormality in the energy efficiency of the equipment itself. In this case, the system determines that the current energy consumption fluctuation is within the normal range and does not trigger an alarm, thus ensuring the stability of the park's operation and maintenance scheduling.

[0070] The original model equations of the ARIMA prediction model satisfy the following relationship:

[0071] ;

[0072] In the formula, Indicates the first The expected energy consumption baseline of the chiller unit at any given time. Represents a constant term. Represents the first-order autoregressive coefficient. Indicates the first The actual power consumption of the chiller unit at all times. Indicates the order of autoregression. Indicates the first Autoregressive coefficient of order, Indicates the first Real-time historical actual power consumption of chiller units This represents the first-order moving average term.

[0073] Substituting the first-order autoregressive coefficients into the original model equation satisfies the following relationship:

[0074] ;

[0075] In the formula, Indicates the first The expected energy consumption baseline of the chiller unit at any given time. Represents a constant term. Represents the first-order autoregressive coefficient. Indicates the current time The drift correction factor Indicates the first The actual power consumption of the chiller unit at all times. Indicates the order of autoregression. Indicates the first Autoregressive coefficient of order, Indicates the first Real-time historical actual power consumption of chiller units This represents the first-order moving average term.

[0076] This invention also provides an energy efficiency monitoring system for industrial parks based on digital twins. For example... Figure 2 As shown, the system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the energy efficiency monitoring method for industrial parks based on digital twins according to the first aspect of the present invention. The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, the settings and functions of which are known in the art and will not be described further here.

[0077] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for monitoring energy efficiency in industrial parks based on digital twins, characterized in that, include: The source thermal shock temperature, source cooling water flow rate, ambient temperature of the cooling pipe network of the return water main in the production workshop, and actual power consumption at the input of the chiller unit's power distribution cabinet are obtained. Set a sliding time window, calculate the average of the historical return water temperature sequence within the sliding time window as the benchmark reference temperature, and calculate the degree of transient thermal power deviation based on the difference between the current return water temperature and the benchmark reference temperature and the current cooling water flow rate, thereby quantifying the heat load disturbance caused by sudden changes in upstream production scheduling. By combining the cooling pipe network parameters and the current cooling water flow rate, the heat transfer delay step and the actual spatial velocity are calculated. Based on the corresponding transient thermal power deviation at the heat transfer delay step and the heat transfer delay step and the normalized actual spatial velocity, the dynamic thermal shock intensity at the current moment is calculated. The drift correction factor at the current moment is calculated based on the dynamic thermal shock intensity. The first-order autoregressive coefficient of the ARIMA model is adjusted by the drift correction factor. The adjusted first-order autoregressive coefficient is substituted into the ARIMA model to obtain the corrected prediction model. The expected energy consumption baseline of the chiller unit at the next moment is predicted. The residual between the actual power consumption of the chiller unit and the expected energy consumption baseline is calculated, and the operating status of the chiller unit is determined. The adjusted first-order autoregressive coefficients satisfy the following relationship: In the formula, This represents the adjusted first-order autoregressive coefficient; This represents the first-order autoregressive coefficients obtained by fitting based on the stationarity assumption using traditional methods; Indicates the current time The drift correction factor; The heat flow transport delay step is calculated as follows: The cooling network parameters include the total length and cross-sectional area of ​​the pipes from the front-end thermal node to the inlet of the terminal chiller unit. The ratio of the current cooling water flow rate to the pipe cross-sectional area is used as the actual spatial velocity. The ratio of the total pipe length to the actual spatial velocity is rounded down to obtain the heat flow transmission delay step.

2. The method for monitoring energy efficiency in industrial parks based on digital twins according to claim 1, characterized in that, The calculation method for the degree of deviation of the transient thermal power is as follows: Calculate the difference between the current return water temperature and the reference temperature, multiply the difference by the current cooling water flow rate, and then calculate the degree of transient thermal power deviation using the hyperbolic tangent function.

3. The method for monitoring energy efficiency in industrial parks based on digital twins according to claim 1, characterized in that, The dynamic thermal shock intensity is calculated as follows: The actual spatial velocity at the current moment is normalized, and the product of the normalized actual spatial velocity and the degree of deviation of the transient thermal power is taken as the dynamic thermal shock intensity at the current moment.

4. The method for monitoring energy efficiency in industrial parks based on digital twins according to claim 1, characterized in that, The drift correction factor is calculated as follows: The difference between the current dynamic thermal shock intensity and the mean dynamic thermal shock intensity within the corresponding sliding time window is taken as the dynamic thermal shock deviation; the average absolute value of the dynamic thermal shock intensity at each historical moment within the sliding time window is taken as the historical average thermal shock intensity, and the sum of the historical average thermal shock intensity and the minimum hyperparameter is taken as the denominator; the dynamic thermal shock deviation is divided by the denominator to obtain the drift correction factor at the current moment.

5. The method for monitoring energy efficiency in industrial parks based on digital twins according to claim 1, characterized in that, The steps to determine the operating status of a chiller unit include: The ARIMA prediction model is updated based on the adjusted first-order autoregressive coefficients to obtain the expected energy consumption baseline of the chiller unit at the current moment. The difference between the actual power consumption of the chiller unit at the current moment and the expected energy consumption baseline is taken as the residual at the current moment. The standard deviation of the historical residual sequence formed by the residuals of the chiller units at each historical moment within the corresponding sliding time window at the current moment is calculated. If the residual at the current moment is greater than 3 times the standard deviation, the system determines that the chiller unit has significantly high energy consumption behavior and outputs an inefficiency alarm. Conversely, if the residual at the current moment is less than or equal to 3 times the standard deviation, the system determines that the current energy consumption fluctuation is within the normal range and does not trigger an alarm.

6. An industrial park energy efficiency monitoring system based on digital twins, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the energy efficiency monitoring method for industrial parks based on digital twins according to any one of claims 1-5.

Images