TWI414734B - Dynamic modeling method, monitoring method for chilling system and device for monitoring chilling system - Google Patents

Abstract

A dynamic modelling method for a chilling system comprises the steps of: calculating a KPI (Key Performance Index) value; obtaining a plurality of control points determining a representative synthetic curve dynamic performance model; calculating a search direction vector for the control pointes wherein each control point is adjusted independently by a corresponding element of the search direction vector, and a computed synthetic curve dynamic performance model may move closer to the KPI value relative to the representative synthetic curve dynamic performance model; providing a movement variable wherein each control point is adjusted by the product of the search direction vector and the movement variable, thereby obtaining a synthetic curve dynamic performance model approximating the KPI value; and determining the movement variable in each iteration according to a technique for finding the extremum of a unimodal function.

Description

Ice water host dynamic characteristic model establishing method, ice water host monitoring method and ice water host monitoring device

The invention relates to a method for establishing a characteristic model of an ice water host.

The air-conditioning ice water system used in large commercial buildings and factories is mainly composed of ice water main engine, ice water pump, cooling water pump, and cooling water tower. The energy demand of the ice water main unit is the highest. For example, a 12-inch wafer fab with a total installed refrigeration tonnage (RT) of 21,000, for example, produces about 26% of the total electricity consumption of air-conditioning ice water, and the ice water host accounts for the total electricity consumption. 50%. Therefore, if you ensure that the ice water main unit operates at the highest efficiency range, you can reduce the considerable energy loss.

The chiller host characteristic curves provided by the manufacturer are generally obtained by direct testing on a non-physical installation site. Due to the influence of the piping system and the operation mode, operating with the characteristic curve provided by the manufacturer often makes it impossible for the ice water main unit to operate at the highest efficiency range, resulting in a decrease in energy efficiency. Therefore, after the ice water main unit is installed, the ice water host characteristic curve will be re-established through the actual operation data of the site.

At present, the establishment of the ice water host characteristic curve can use two models: regression model and artificial intelligence model. However, these two models have the following disadvantages: (1) A certain amount of operational data needs to be prepared, and too little operational data will not be able to establish the model. For example: if it is a 3rd-order polynomial, at least 4 different RT operation data are needed to establish a curve; (2) such as YH Song et al. (Young-hak Song, Yasunori Akashi, Jurng-Jae Yee, 2007, "Energy performance of a cooling plant system using the inverter chiller for industrial building," Energy and Building, Vol. 39, Issue 3, Pages 289-297), different cooling water return water temperature, need to use different orders The regression model causes the difficulty of model building. (3) When the performance curve has small regional characteristics (such as two valleys), the order of the regression model will increase rapidly, resulting in a large increase in the amount of calculation and data demand; and the use of artificial intelligence models also requires more Calculate data to simulate small area features.

In addition to the above problems, the use of regression models and artificial intelligence models, when the ice water host feature model is established, requires a long training data. On the other hand, the ice water main engine needs to be repaired after a long period of operation. However, after the ice water main unit is repaired and maintained, its actual main engine running characteristics will differ from the characteristics at the time of modeling. Under this circumstance, the original characteristic curve needs to be adjusted, and the characteristic curve established by using the regression model and the artificial intelligence model will once again be trained for a long time, and the characteristic curve cannot be directly and dynamically adjusted locally. Corrected.

According to the above problem, an embodiment of the present invention provides a method for establishing a dynamic characteristic model of an ice water host, which comprises the following steps: acquiring an ice water return water temperature, an ice water water discharge temperature, and a cooling water return water temperature of an ice water host. a flow rate and a power consumption amount; according to the ice water return water temperature, the ice water outlet water temperature, the flow rate and the electricity consumption amount, calculating an actual operation efficiency value; obtaining a plurality of control points, wherein the control points represent a synthetic curve dynamic characteristic model; using a disturbance amount, calculating a corresponding element value of each control point in a moving direction vector, wherein each of the control points is individually adjusted by a corresponding element value, and one of the obtained calculated curves is calculated The dynamic characteristic model is closer to the actual operational efficiency value than the synthetic curve dynamic characteristic model; a moving variable is provided, and the control points are adjusted by the product of the moving variable and the moving direction vector, thereby obtaining the approximate operational efficiency One of the values of the composite curve dynamics model; and in each iterative calculation, the moving variable is determined according to a variable minimum search method Value.

Another embodiment of the present invention provides a method for monitoring an ice water host, comprising the steps of: obtaining an ice water return water temperature, an ice water water outlet temperature, a cooling water return water temperature, a flow rate, and a power consumption; The ice water return water temperature, the ice water water discharge temperature, the flow rate and the electricity consumption amount, calculate a partial load ratio and an actual operation efficiency value; obtain a plurality of control points, wherein the control points determine a representative synthetic curve dynamic characteristic model Using a disturbance amount, calculating a corresponding one of the element values in one of the movement direction vectors of each of the control points, wherein each of the control points is individually adjusted by the corresponding element value, and one of the obtained calculations is a synthetic curve dynamic characteristic model compared to the synthesis The curve dynamic characteristic model is closer to the actual operating efficiency value; a moving variable is provided, and the control points are adjusted by the product of the moving variable and the moving direction vector, thereby obtaining a synthetic curve dynamic that approximates the actual operating efficiency value Characteristic model; in each iterative calculation, the moving variable is determined according to a variable minimum search method; and when an iteration abort element is satisfied Generating a new dynamic characteristic model representing composite curve.

Another embodiment of the present invention provides an ice water host monitoring device including a capture device, an arithmetic device, and a display device. The pick-up device can obtain information on the operation of an ice water host. The computing device may calculate an actual operating efficiency value of one of the ice water hosts according to the operation information, and adjust a plurality of control points such that a composite curve dynamic characteristic model determined by the plurality of control points approaches the actual operating efficiency value . The display device displays a synthetic curve dynamic characteristic model representing one of the operational information.

The ice water host dynamic characteristic model established by the synthetic curve can be dynamically updated in the area of the required change, and the ice water host characteristic model can be dynamically established. In addition, if the control point is added more accurately to describe the characteristics of the ice water host, the increase of the calculation cost will not be caused by the unrestricted increase of the power of the curve.

1 shows a connection diagram of a chilled water host monitoring system 1 according to an embodiment of the present invention. 2 shows a schematic diagram of an ice water host monitoring system 1 according to an embodiment of the present invention. Referring to FIG. 1 and FIG. 2, the ice water host monitoring system 1 can be connected to an ice water host or a factory monitoring system (FMCS) 2, through the sensor or factory database on the ice water host 2, to capture The device 11 acquires operation information related to the ice water host, and the information includes the power consumption W of the ice water host 2, the ice water return water temperature T _chi , the ice water water temperature T _cho , the cooling water return water temperature T _cwi , and the ice water Operation information such as traffic Q _cho . Calculate the actual ice water main operation efficiency (KPI _real ), output frozen tons (RT) and partial load ratio PLR _current from the ice water host dynamics through the information obtained, and pass the B-spline when cooling the return water temperature The KPI _model value obtained by the characteristic model is then displayed on the visual interface connected to a display device 14 to assist the operator in the operation of the chilled water host 2.

The computing device 12 can use the foregoing operational information, calculate the actual ice water host operating efficiency (KPI _real ), the output freezing ton (RT) and the partial load ratio PLR _current , and the actual operating performance data can be borrowed. It is represented by the following formula (1), formula (2), and formula (3):

RT = ( T _chi - T _cho ) × Q _ch (2)

Among them, KPI _{real is} the actual operating efficiency value, RT is the frozen tonnage input by the ice water main unit, T _chi is the ice water return water temperature, T _cho is the ice water outlet temperature, Q _cho is the flow rate, and W is the electricity consumption. The PLR (Partial Loading Ratio) is a partial load ratio, RT _specific is the rated freezing ton of the ice water host 2, and T _cwi is the cooling water return water temperature.

The computing device also includes a calculation to establish a B-spline dynamic characteristic model of the ice water host which can represent the aforementioned operation information, a definition of the B-spline curve and a correction method of the optimum curve, and the single cooling water return water temperature is represented by a B-spline curve. The ice water host dynamic characteristic model can also express its characteristic model as a B-spline surface to express its cooling water return water temperature parameters.

The invention can use any synthetic curve to establish a dynamic curve model of the synthetic curve. In the embodiment of the present invention, the model is built by a B-spine curve or a curved surface, and other synthetic curves such as a Bezier curve can also be built in a similar method flow, so this specification The Bezier curve modeling process is no longer described.

The completed ice water host characteristic model and its related parameter data (the rated freezing tons of the ice water host) are stored in a storage device 15 (for example, a hard disk or a tape drive, etc.) and displayed on the display device 14 On, for operation.

As shown in FIG. 3, the computing device 12 is also configured to utilize the actual operational performance data in the storage device 15 and the historically established ice water host dynamics model to obtain the ice water host dynamic characteristic model by trend statistical analysis. The trend is changed, and when the variation of the operational efficiency value exceeds a threshold, the operator is warned by the warning device 13. In addition, the computing device 12 is also constructed under the same operating conditions. By comparing the efficiency value distributions of the two periods before and after, the trend of the equipment operating efficiency data is obtained, and whether the performance is obtained through the operating performance and the characteristic curve of the device. In the high-efficiency operation section, it is possible to operate in a high-efficiency section by changing the relevant operating parameters to achieve energy saving.

In an embodiment, the capture device 11 can include a plurality of input ports, wherein the inputs are correspondingly connected to sensors on the ice water host 2, or to their factory operation system or central monitoring system. The computing device 12 can include a central processing unit and a memory, wherein the central processing unit, the memory, and the input ports can utilize a bus bar for data transfer. Display device 14 can include a screen. The storage device 15 can include a hard disk.

4 is a flow chart showing a method for establishing a dynamic model of a glacial water host according to an embodiment of the present invention.

In step S40, first, an ice water return water temperature (T _chi ), an ice water outlet water temperature (T _cho ), a cooling water return water temperature (T _cwi ), and an ice water flow rate (Q _cho ) of the ice water host 2 are obtained. And operating information such as a power consumption (W), and calculating the freezing tonnage and partial load ratio (PLR) of the ice water host 2, wherein the temperature, water flow and power consumption are directly measured, or Obtained in the factory monitoring system.

In step S41, an actual operational efficiency value (KPI _real ) is calculated according to the aforementioned formulas (1) and (2).

In step S42, a plurality of control points are defined or obtained, wherein the plurality of control points determine that one of the running characteristics of the ice water host represents a B-spline dynamic characteristic model. When the icy water host dynamic characteristic model (KPI _model = Bspline (PLR) at a specific cooling water return water temperature) indicating the characteristic curve of the glazed water main unit 2 is established, the following formulas (4) and (5) are utilized.

Where n is the number of control points, k is the power of the B-spline curve function, P _i is the control point of the B-spline curve function, and N _i,k is the mixed function of B-spline, which can be expressed as .

Since the characteristics of the B-spline curve are that the function of the curve shape can be locally controlled and changed, and the control points of the curve are increased, the curve equations do not increase accordingly. Due to the characteristics of these two items, when the ice water host characteristic model is established by the B-spline curve, the curve can be effectively updated in the area of the required change, and the ice water host characteristic model can be dynamically established. In addition, if the characteristics of the ice water host are more accurately described and the control points of the curve are increased, the increase of the calculation cost will not be caused by the unrestricted increase of the power of the curve. The B-spline curve allows the user to set or insert a curve control point at any position of interest to make the curve the desired curve shape.

If the cooling water return water temperature is considered, the ice water host dynamic characteristic model (KPI _model = Bspline (T _cwi , PLR)) must be established by the following formula (8).

The B-spline dynamic characteristic model can be the dynamic characteristic model of the ice water host in use, and the B-spline dynamic characteristic model can also be an initial dynamic characteristic model calculated by using a plurality of initial control points whose values are fixed. If the B-spline dynamic characteristic model is the initial dynamic characteristic model, after determining the number of control points, the control point is given a fixed initial value, for example: 1. If the control points are combined into a Performance Surface of Matrix ( PSM ) representation, then PSM _{1 × n} = [1, 1, ..., 1] _{1 × n} , where n is the number of control points, at this time KPI _model = 1.

In step S43, an optimum moving direction of one of the control points is calculated, wherein the optimum moving direction of all the control points can be combined into a moving direction vector s , and s can be calculated by the following formulas (9) and (10).

f =| COP _{mod el} - COP _real | (9)

The calculation formula of the moving direction of each control point (10) knows that the corresponding moving direction vector element value of each control point is f/ x _i , that is, the i control point performs a disturbance amount dx, the remaining control points maintain the original value, and the partial load ratio (PLR) condition obtained in step S40 is substituted into the calculated B-spline curve constructed by the new control point. Dynamic characteristic model, obtain the KPI _model value under this new control point, and calculate f=f(x)-f(x+ x) the value of the function below, if the value is positive, indicating the amount of disturbance, so that the efficiency value obtained by the B-spline dynamic characteristic model of the new control point is closer to the actual operational efficiency calculated in step S41. The value, that is, the B-spline curve characteristic also tends to be the actual ice water host operating characteristic, and the moving direction caused by this disturbance is the correct direction.

In steps S44 and S45, a movement variable α is provided. The control point is adjusted by the product of the moving variable α and the optimum moving direction vector, thereby obtaining a B-spline dynamic characteristic model that approximates the actual operating efficiency value. The adjustment of the control point is an iterative manner in which a new B-spline dynamics model representative of the actual operational efficiency value calculated at step S41 is gradually obtained. If expressed mathematically, it can be expressed by the following formula (11).

PSM _{k +1} = PSM _k + α × s (11)

Where k is the iteration number. The movement variable α is a step length for adjusting the control point, which can be searched by a variable minimum value method such as Newton method, Golden Section Search, nerve-like algorithm, gene calculus. Method or Fibonacci Search method (Fibonacci Search) and other methods.

Referring to Fig. 5, in order to obtain a new B-spline dynamic characteristic model representing the actual operational efficiency value calculated in step S41, the f value of equation (9) must be minimized. To this end, the single variable minimum value search method can be used to search for the movement variable α which minimizes the f value. If the gold cutting method is used, an interval [a, b] is determined first, where a and b are the search values of the previous iteration. Then, by selecting a new moving variable α _i in the interval of (ba) × τ (=0.382) in Fig. 5, the interval [a, b] can be reduced by 1-τ times. By repeating the foregoing steps repeatedly, KPI _{real which} minimizes the f value of the formula (9) can be obtained.

In step S46, a plurality of adjustment control points are calculated according to formula (11).

In step S47, an adjusted B-spline dynamic characteristic model is calculated according to the plurality of adjustment control points obtained in step S46.

In step S48, a first difference value between the adjusted B-spline dynamic characteristic model of step S47 and the actual operational efficiency value (KPI _real ) calculated in step S41 is calculated according to formula (9).

In step S49, according to formula (9), a second difference value between a previously adjusted B-spline dynamic characteristic model and the actual operational efficiency value is calculated.

In step S50, it is an iterative suspension requirement. The definition of the iterative suspension requirement may be to compare the first difference value with the second difference value. If the quantity difference value is less than a set value, the iterative process is aborted, or may be defined as If the difference value is less than a set value, the iterative process is aborted, or may be defined as the number of iterations. When the number of iterations is greater than a set value, the iterative process is aborted. When the iterative suspension requirement is not reached, it indicates that the optimized B-spline dynamic characteristic model has not been obtained, and then returns to step S45 to calculate a new movement variable α.

In step S51, when the iterative suspension requirement is satisfied, the adjusted B-spline dynamic characteristic model calculated in step S47 is listed as a new representative B-spline dynamic characteristic model.

In an embodiment, the B-spline dynamic characteristic model is an initial dynamic characteristic model, and after determining the number of control points, a certain value of the control point is given, for example: 1. If the control points are combined into a Performance Surface of Matrix ( PSM ) representation, then =[1,1,...,1] _1xn , where n is the number of control points, at which point KPI _model =1. Then, according to the steps S40 to S50 described above, a representative B-spline dynamic characteristic model is calculated.

FIG. 6 is a flow chart showing a method for monitoring an ice water host according to an embodiment of the present invention. In step S60, first, an ice water return water temperature (T _chi ), an ice water outlet water temperature (T _cho ), a cooling water return water temperature (T _cwi ), and an ice water flow rate (Q _cho ) of the ice water host 2 are obtained. And a power consumption (W), and calculate the freezing tonnage and partial load ratio (PLR) of the ice water host 2, wherein the aforementioned temperature, water flow and power consumption are directly measured, or from the factory monitoring Obtained in the system.

In step S61, an actual operational efficiency value (KPI _real ) is calculated according to the aforementioned formulas (1) and (2).

In step S62, a plurality of control points are obtained, wherein the plurality of control points determine that one of the running characteristics of the ice water host represents a B-spline dynamic characteristic model KPI _model = Bspline (PLR) or KPI _model = Bspline (T _cwi , PLR).

In step S63, an optimum moving direction of one of the control points is calculated, wherein the optimum moving direction of all the control points can be combined into a vector s , and s can be calculated by the following formulas (9) and (10).

f =| COP _{mod el} - COP _real | (9)

The calculation formula of the moving direction of each control point (10) knows that the corresponding moving direction vector element value of each control point is f/ x _i , that is, the i control point performs a disturbance amount dx, the remaining control points maintain the original value, and the partial load ratio (PLR) condition obtained in step S60 is substituted into the B-spline curve dynamic constructed by the new control point. Characteristic model, obtain the KPI _model value under this new control point, and calculate f=f(x)-f(x+ x) the value of the function below, if the value is positive, indicating the amount of disturbance, so that the efficiency value obtained by the B-spline dynamic characteristic model of the new control point is closer to the actual operational efficiency calculated in step S61. The value, that is, the B-spline curve characteristic also tends to be the actual ice water host operating characteristic, and the moving direction caused by this disturbance is the correct direction.

In steps S64 and S65, a movement variable α is provided. As shown in the formula (11), the control point is adjusted by the product value of the movement variable α and the optimum moving direction vector of each control point, thereby obtaining a B-spline dynamic characteristic model which approximates one of the actual operational efficiency values. The adjustment of the control points is in an iterative manner, gradually obtaining a new B-spline dynamics model representative of the actual operational efficiency values calculated at step S61.

In step S66, a plurality of adjustment control points are calculated according to formula (11).

In step S67, an adjusted B-spline dynamic characteristic model is calculated according to the plurality of adjustment control points obtained in step S66.

In step S68, a first difference value between the adjusted B-spline dynamic characteristic model of step S67 and the actual operational efficiency value (KPI _real ) calculated in step S61 is calculated according to formula (9).

In step S69, according to formula (9), a second difference value between a previously adjusted B-spline dynamic characteristic model and the actual operational efficiency value is calculated.

In step S70, it is an iterative suspension requirement. The definition of the iterative suspension requirement may be comparing the first difference value with the second difference value. If the quantity difference value is less than a set value, the iteration process is aborted, or may be defined as If the difference value is less than a set value, the iterative process is aborted, or may be defined as the number of iterations. When the number of iterations is greater than a set value, the iterative process is aborted. When the iterative suspension requirement is not reached, it indicates that the optimized B-spline dynamic characteristic model has not been obtained, and the process returns to step S65 to calculate a new movement variable α.

In step S71, when the iterative suspension requirement is satisfied, the adjusted B-spline dynamic characteristic model calculated in step S67 is listed as a new representative B-spline dynamic characteristic model.

In step S72, based on the comparison result between the new representative B-spline dynamic characteristic model and the original representative B-spline dynamic characteristic model, it is determined whether or not an abnormal warning is raised. In an embodiment, a warning is issued when the difference between the new representative B-spline dynamics model and the original representative B-spline dynamics model is greater than a threshold.

In another embodiment, the ice water host monitoring system 1 can further calculate a plurality of old representative B-spline dynamic characteristic models and a new representative B-spline dynamic characteristic model, a plurality of theories under a partial load ratio An operational efficiency value; and a variation of the plurality of theoretical operational efficiency values, and a warning if the variance is greater than a threshold.

In another embodiment, the ice water host monitoring system 1 can count the actual operating efficiency values of the plurality of pens at a partial load ratio, and generate a warning if the variation of the actual operating efficiency values is greater than a threshold.

The new representative B-spline dynamic characteristic model can be displayed on the display device 14 with the dynamic characteristic model and the original representative B-spline dynamic characteristic model for the operator to monitor the trend of the dynamic characteristic model.

The technical content and technical features of the present disclosure have been disclosed above, but those skilled in the art can still make various substitutions and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of the present disclosure is not to be construed as limited by the scope of

1. . . Ice water host monitoring system

2. . . Ice water host

11. . . Pickup device

12. . . Arithmetic device

13. . . Warning device

14. . . Display device

15. . . Storage device

S40~S51. . . Process step

S60~S72. . . Process step

1 is a schematic diagram showing the connection of an ice water host monitoring system according to an embodiment of the present invention;

2 is a schematic view showing an ice water host monitoring system according to an embodiment of the present invention;

FIG. 3 illustrates a change trend of the dynamic characteristics model of the ice water host and an efficiency value distribution of the front and rear time periods according to an embodiment of the present invention;

4 is a flow chart showing a method for establishing a dynamic characteristic model of an ice water host according to an embodiment of the present invention;

Figure 5 illustrates the search for minimum movement variables by gold cutting; and

FIG. 6 is a flow chart showing a method for monitoring an ice water host according to an embodiment of the present invention.

S40~S51. . . Process step

Claims (21)

An ice water host dynamic characteristic model establishing method comprises the following steps: obtaining an ice water return water temperature, an ice water water outlet temperature, a cooling water return water temperature, a flow rate and a power consumption; Ice water return water temperature, the ice water water discharge temperature, the cooling water return water temperature, the flow rate and the electricity consumption amount, calculate a part load ratio and an actual operation efficiency value; obtain a plurality of control points, wherein the control Point determining a representative synthetic curve dynamic characteristic model; using a disturbance amount, calculating a corresponding element value of each of the control points in a moving direction vector, wherein each of the control points is individually adjusted by the corresponding element value, and the obtained A calculated synthetic curve dynamic characteristic model is closer to the actual operational efficiency value than the original synthetic curve dynamic characteristic model; a moving variable is provided, and the control points are adjusted by the product of the moving variable and the moving direction vector, thereby obtaining an approximation One of the actual operational efficiency values is a composite curve dynamics model; and in each iterative calculation, the movement is determined according to a variable minimum search method The value of the number. The ice water host dynamic characteristic model establishing method according to claim 1, further comprising the steps of: calculating a plurality of adjustment control points; calculating an adjusted synthetic curve dynamic characteristic model by using the adjustment control points; comparing the actual The operating efficiency value and the adjusted synthetic curve dynamic The sexual model; and when an iterative suspension requirement is satisfied, the adjusted synthetic curve dynamic characteristic model is used as a new representative synthetic curve dynamic characteristic model. The ice water host dynamic characteristic model establishing method according to claim 1, wherein the decision of the moving variable is obtained by the variable minimum value searching method. The method for establishing a chilled water host dynamic characteristic model according to claim 3, wherein the variable minimum search method comprises a golden cut search method, a Newton method, a neural-like algorithm, a gene algorithm or a Fibonacci search method. The ice water host dynamic characteristic model establishing method according to claim 1, wherein the moving direction vector is determined according to the following steps: adjusting the control point by the disturbance amount; calculating a corresponding dynamic model of the calculated synthetic curve; Comparing with the representative dynamic curve dynamic characteristic model, if the calculated synthetic curve dynamic characteristic model is close to the actual operational efficiency value, the corresponding calculation between the calculated synthetic curve dynamic characteristic model and the representative synthetic curve dynamic characteristic model is calculated. And a ratio between the difference value of the partial load ratio and the disturbance amount, wherein the ratio is the corresponding element value of the control point in the moving direction vector. The ice water host dynamic characteristic model establishing method according to claim 1, wherein the representative synthetic curve dynamic characteristic model is calculated by using a plurality of fixed initial value control points. The ice water host dynamic characteristic model establishing method according to claim 1, wherein the actual operating efficiency value is calculated by the following formula: KPI _{real is} the actual operating efficiency value, T _{chi is} the ice water return water temperature, T _{cho is} the ice water water outlet temperature, O _{cho is} the flow rate, and W is the power consumption. An ice water host monitoring method comprises the steps of: obtaining an ice water return water temperature, an ice water water outlet temperature, a cooling water return water temperature, a flow rate and a power consumption; according to the ice water back a water temperature, an ice water outlet temperature, a cooling water return water temperature, the flow rate and the electricity consumption amount, calculating a partial load ratio and an actual operating efficiency value; obtaining a plurality of control points, wherein the control points determine one Representing a synthetic curve dynamic characteristic model; using a disturbance amount, calculating a corresponding element value in one of the movement direction vectors of each of the control points, wherein each of the control points is individually adjusted by the corresponding element value, and one of the obtained calculations is synthesized The curve dynamic characteristic model is closer to the actual operating efficiency value than the original synthetic curve dynamic characteristic model; a moving variable is provided, and the control points are adjusted by the product of the moving variable and the moving direction vector, thereby obtaining the approximation of the actual operation One of the efficiency values is a composite curve dynamic characteristic model; in each iterative calculation, the moving variable is determined according to a variable minimum search method; Abort requirements satisfied, generate a new representative synthesis dynamic characteristic model curve. The ice water host monitoring method according to claim 8, further comprising: determining a step of issuing an abnormality warning according to a comparison result between the new representative synthetic curve dynamic characteristic model and the representative synthetic curve dynamic characteristic model. The ice water host monitoring method according to claim 8, further comprising the steps of: obtaining a plurality of old representative synthetic curve dynamic characteristic models; calculating the new representative synthetic curve dynamic characteristic model and the plurality of old representative synthetic curve dynamics The characteristic model is a plurality of theoretical operational efficiency values at a portion of the load ratio; and a variation of the plurality of theoretical operational efficiency values is calculated, and a warning is generated based on the variation. The ice water host monitoring method according to claim 8, further comprising the steps of: calculating a plurality of adjustment control points; calculating an adjusted synthetic curve dynamic characteristic model by using the adjustment control points; and comparing the actual operation efficiency values; And the adjusted synthetic curve dynamic characteristic model; and when the iterative suspension requirement is satisfied, the adjusted synthetic curve dynamic characteristic model is used as a new representative synthetic dynamic characteristic model. The chilled water host monitoring method according to claim 8, wherein the decision of the movement variable is obtained by the variable minimum value search method. The ice water host monitoring method according to claim 12, wherein the variable minimum search method comprises a gold cut search method, a Newton method, a neural algorithm, Gene algorithm or Fibonacci search method. The ice water host monitoring method according to claim 8, wherein the moving direction vector is determined according to the following steps: adjusting the control point by the disturbance amount; calculating a dynamic model of the calculated synthetic curve; and Synthesizing curve dynamic characteristic model comparison, if the calculated synthetic curve dynamic characteristic model is close to the actual operating efficiency value, the corresponding calculated synthetic curve dynamic characteristic model and the representative synthetic curve dynamic characteristic model are calculated in the partial load And a ratio between the difference value and the disturbance amount value, wherein the ratio is the corresponding element value of the control point in the moving direction vector. The chilled water host monitoring method according to claim 8, wherein the representative synthetic curve dynamic characteristic model is calculated by using a plurality of fixed initial value control points. The ice water host monitoring method according to claim 8, further comprising the steps of: counting the actual operating efficiency value of the plurality of pens at the partial load ratio; and if the actual operating efficiency value is greater than one A warning is generated when the threshold is negative. The ice water host monitoring method according to claim 8, wherein the actual operating efficiency value is calculated by the following formula: KPI _{real is} the actual operating efficiency value, T _{chi is} the ice water return water temperature, T _{cho is} the ice water water outlet temperature, Q _{cho is} the flow rate, and W is the power consumption. An ice water host monitoring device comprises: a capturing device for acquiring operation information of an ice water host; and an computing device configured to calculate an actual operating efficiency value of the ice water host according to the operation information, and adjust a plurality of control points such that a composite curve dynamic characteristic model determined by the plurality of control points approaches the actual operational efficiency value; and a display device for displaying a dynamic characteristic model representing the synthetic curve representing one of the operational information . The ice water host monitoring device according to claim 18, wherein the capturing device is connected to a sensor, a factory operation system or a central monitoring system on the ice water host. The ice water host monitoring device according to claim 18, further comprising a warning device, wherein the computing device calculates a variation amount of the operational efficiency value, and when the variation of the operational efficiency value exceeds a threshold value, The warning device will issue a warning. The chilled water host monitoring device of claim 18, wherein the synthetic curve dynamics model comprises a B-spine or Bezier curve model.