Water pump flow simulation calculation method based on nonlinear least square modeling

Abstract

The invention relates to the technical field of simulation calculation methods, in particular to a water pump flow simulation calculation method based on nonlinear least square modeling. Observing and counting the power p and lift h of the water pump and the flow f of the corresponding pipeline flowmeter at the same time; forming a data set matrix based on data of observation statistics; generating a simulation calculation model fitting function based on the data set matrix and observation statistical data; building a calculation and analysis model based on water pump flow monitoring; the built calculation and analysis model is combined with the water pump control flow simulation calculation model, flow calculation and analysis based on water pump power and lift are achieved, accurate estimation of pipeline flow is achieved, dependence on a flow meter is reduced, and economical efficiency and flexibility of flow monitoring are improved. Data support is provided for optimizing the operation condition of the water pump, and debugging efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of simulation calculation methods, and in particular to a water pump flow simulation calculation method based on nonlinear least squares modeling. Background Art

[0002] In the operation of secondary water supply systems and industrial pumping stations, flow measurement is a key link in evaluating system performance and ensuring stable operation. Flow measurement not only helps to understand the system's water supply capacity and operating status, but also can promptly detect potential faults and abnormal conditions, thereby ensuring the safe and stable operation of the system.

[0003] Currently, flow measurement in secondary water supply systems and industrial pumping stations usually relies on flow meters, but the installation cost of flow meters is high, and the layout and maintenance of flow meters are difficult in complex pipe network environments. Summary of the Invention

[0004] The purpose of the present invention is to provide a water pump flow simulation calculation method based on nonlinear least squares modeling. By collecting the operating power and head data of the water pump and combining it with the pipeline characteristics, a three-dimensional statistical model is established to achieve accurate estimation of the pipeline flow. This method can reduce dependence on flow meters, improve the economy and flexibility of flow monitoring, and at the same time provide data support for optimizing the operating conditions of the water pump during the water pump debugging process, thereby improving debugging efficiency.

[0005] To achieve the above object, the present invention provides a water pump flow simulation calculation method based on nonlinear least squares modeling, comprising the following steps:

[0006] Observe and count the power p and lift h of the water pump at the same time, as well as the flow rate f of the corresponding pipeline flowmeter, and construct a data set matrix based on the observed and counted data;

[0007] Generating a simulation calculation model fitting function based on the data set matrix and observation statistical data;

[0008] Build a calculation and analysis model based on water pump flow monitoring;

[0009] The constructed calculation and analysis model is combined with the water pump control flow simulation calculation model to realize flow calculation and analysis based on water pump power and head.

[0010] The specific method of forming the data set matrix is ​​as follows:

[0011] Based on the actual demand statistics of the corresponding data of group i at different times, the pump power p of group i is obtained i , pump head h i And the water pipe flow counter value f i , where: i = {1, 2, ..., n};

[0012] The following matrix is ​​constructed based on the obtained n sets of pump power, pump head and pipe flow rate count values:

[0013] The dataset matrix mainly includes:

[0014] Water pump dataset matrix:

[0015] Traffic matrix:

[0016] The specific construction method of the simulation calculation model fitting function is as follows:

[0017] Based on the constructed data set matrix, a binary quadratic nonlinear model is preset, and its function is: f i =c1p i 2 +c2h i 2 +c3p i h i +c4p i +c5h i +c6;

[0018] The model matrix formula based on the constructed data set matrix combined with the preset binary quadratic nonlinear model is: F=DC, where the parameter matrix C to be determined is:

[0019] Solving the parameter matrix C in the model matrix formula;

[0020] Substitute the solved parameter matrix C into the preset binary quadratic nonlinear model to obtain the simulation calculation model fitting function f=c1p 2 +c2h 2 +c3ph+c4p+c5h+c6.

[0021] The specific construction method of the calculation and analysis model based on water pump flow monitoring is as follows:

[0022] Obtain the collected water pump power p and head h, as well as the corresponding pipeline flow meter flow f data;

[0023] The collected data is cleaned by removing outliers, and then constructed into a matrix data set and stored in the local database;

[0024] Based on the matrix data set constructed above, a computational analysis model based on water pump flow monitoring is established using statistical analysis and multidimensional regression methods.

[0025] The specific method of flow calculation and analysis based on pump power and head is as follows:

[0026] Exporting the computational analysis model, and visualizing the model appearance and error comparison;

[0027] Perform data screening and alerts;

[0028] By screening data with large error values, the accuracy of the simulation calculation model can be improved.

[0029] The specific solution method of the parameter matrix C to be determined is:

[0030] Expand the sum of squared errors: E = ||F-DC|| 2 =(F-DC) T (F-DC), taking the derivative of C at both ends and setting it equal to 0, we get: Simplified to: D T DC=D T F;

[0031] Based on the above simplified formula D T DC=D T F, solve the parameter matrix C to be determined, if D T D is reversible, so we can solve C: C=(D T D) -1 D T F.

[0032] The specific method of comparing the appearance and error of the visualization model is as follows:

[0033] Integrating the computational analysis model into a water pump monitoring system;

[0034] Real-time traffic estimation through embedded edge computing devices;

[0035] The model calculation results are displayed using a visualization program written based on the Matplotlib library and compared with the actual measurement values ​​of the flow meter to analyze the errors in real time.

[0036] The water pump flow simulation calculation method based on nonlinear least squares modeling of the present invention observes and counts the power p and head h of the water pump at the same time, as well as the corresponding pipeline flow meter flow f, and then constructs a data set matrix based on the observed and statistical data. Based on the data set matrix and the observed statistical data, a simulation calculation model fitting function is generated, and a calculation and analysis model based on water pump flow monitoring is built. The built calculation and analysis model is combined with the water pump control flow simulation calculation model. By collecting the operating power and head data of the water pump and combining the pipeline characteristics, a three-dimensional statistical model is established to achieve accurate estimation of the pipeline flow. This method can reduce dependence on the flow meter, improve the economy and flexibility of flow monitoring, and at the same time, provide data support for optimizing the water pump operating conditions during the water pump debugging process, thereby improving debugging efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art.

[0038] Figure 1 The present invention is a flow chart of a water pump flow simulation calculation method based on nonlinear least squares modeling.

[0039] Figure 2 It is the training set data of the present invention.

[0040] Figure 3 It is the validation set data of the present invention.

[0041] Figure 4 The data of the present invention is collected near the lift of 0.4.

[0042] Figure 5 The data of the present invention is collected near the lift of 0.45.

[0043] Figure 6 The data of the present invention is collected near the lift of 0.5.

[0044] Figure 7 The data of the present invention is collected near the lift of 0.55.

[0045] Figure 8 The data of the present invention is collected near the lift of 0.6.

[0046] Figure 9 The data of the present invention is collected near the lift of 0.65.

[0047] Figure 10 The data of the present invention is collected near the lift of 0.7.

[0048] Figure 11 The data of the present invention is collected near the lift of 0.75. DETAILED DESCRIPTION

[0049] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0050] In the description of the present invention, it should be understood that “plurality” means two or more than two, unless otherwise clearly defined.

[0051] See also Figures 1 to 11 , the present invention provides a water pump flow simulation calculation method based on nonlinear least squares modeling: comprising the following steps;

[0052] S1: Observe and count the power p and head h of the water pump at the same time, as well as the flow rate f of the corresponding pipeline flowmeter, and construct a data set matrix based on the observed and statistical data;

[0053] Furthermore, the specific method of forming the data set matrix is:

[0054] Based on the actual demand statistics of n groups of corresponding data at different times, the pump power p i , pump head h i And the water pipe flow counter value f i , where: i = {1, 2, ..., n};

[0055] Specifically, during the data collection process, the operating power P and output head H of the water pump are measured synchronously, and the corresponding flow rate F is recorded by the pipeline flow meter. The specific observation method is as follows:

[0056] Measuring equipment: Pump operating power P: Use ABB inverter visual CNC module (such as ABB_ACS410) to record pump power data in real time;

[0057] Output head H: The inlet and outlet pressures are measured by the pump inlet and outlet pressure sensors, and the head data is converted;

[0058] Pipeline flow F: Use ultrasonic flowmeter or electromagnetic flowmeter, installed on the pipeline node of the pipeline network for measurement.

[0059] Data collection and recording: PLC is used for real-time synchronous data collection to ensure the time consistency of power, head and flow data;

[0060] The sampling frequency was set to 10 Hz (10 times per second) to ensure the continuity and accuracy of the data;

[0061] The data is stored in the local database and the timestamp is recorded.

[0062] Observation environment: The test experiment is carried out on a dedicated testing platform for secondary water supply equipment to ensure that the water pump is in a stable operating state;

[0063] Different operating conditions (water pump start and stop, different flow requirements) are selected for data collection to cover common operating conditions.

[0064] The data based on observation statistics constitutes a data set matrix;

[0065] Furthermore, the data set matrix mainly includes:

[0066] Water pump dataset matrix:

[0067] Traffic matrix:

[0068] S2: generating a simulation calculation model fitting function based on the data set matrix and observation statistical data;

[0069] Furthermore, the specific construction method of the simulation calculation model fitting function is:

[0070] S201: Based on the constructed data set matrix, a binary quadratic nonlinear model is preset, and its function is: f i =c1p i 2 +c2h i 2 +c3p i h i +c4p i +c5h i +c6;

[0071] S202: Based on the constructed data set matrix and the preset binary quadratic nonlinear model, a model matrix formula is constructed: F=DC, where the parameter matrix C to be determined is:

[0072] S203: Solving the parameter matrix C in the model matrix formula;

[0073] S204: Substitute the solved parameter matrix C into the preset binary quadratic nonlinear model to obtain the simulation calculation model fitting function f=c1p 2 +c2h 2 +c3ph+c4p+c5h+c6.

[0074] Furthermore, the specific solution method of the parameter matrix C to be determined is:

[0075] S2031: Expanded error sum of squares: E = ||F-DC||2 =(F-DC) T (F-DC), taking the derivative of C at both ends and setting it equal to 0, we get: Simplified to: D T DC=D T F;

[0076] S2032: Based on the above simplified formula D T DC=D T F, solve the parameter matrix C to be determined, if D T D is reversible, so we can solve C: C=(D T D) -1 D T F.

[0077] Specifically, the method for constructing the model adopts the nonlinear least square method, which minimizes the sum of squares of the errors between the model prediction value and the observation value E=||F-DC|| 2 To estimate the model parameters, that is, by solving To find the parameters that meet the requirements.

[0078] S3: Build a computational analysis model based on water pump flow monitoring;

[0079] Furthermore, the specific construction method of the calculation and analysis model based on water pump flow monitoring is:

[0080] S301: Obtain the collected data of the water pump power p and head h, and the corresponding pipeline flow meter flow f;

[0081] S302: Cleaning the collected data by removing outliers, and then constructing it into a matrix data set and storing it in a local database;

[0082] S303: Based on the matrix data set constructed above, a calculation and analysis model based on water pump flow monitoring is established using statistical analysis and multidimensional regression methods.

[0083] Specifically, in the pump station operating environment, an ABB inverter visual CNC module, pressure sensor and flow meter are installed to collect real-time data on pump power P, head H and pipeline flow F. The collected data is cleaned and eliminated of outliers, and then constructed into a matrix data set and stored in the local database. Subsequently, based on this data set, a flow calculation model is established using statistical analysis and multidimensional regression methods.

[0084] S4: Utilize the constructed calculation and analysis model in combination with the water pump control flow simulation calculation model to realize flow calculation and analysis based on water pump power and head.

[0085] Furthermore, the specific method of flow calculation and analysis based on pump power and head is as follows:

[0086] S401: Exporting the computational analysis model, and visualizing the model appearance and error comparison;

[0087] S402: Perform data screening and warning;

[0088] S403: Improving the accuracy of the simulation calculation model by screening data with large error values.

[0089] Furthermore, the specific method of comparing the appearance and error of the visualization model is as follows:

[0090] S4011: Integrating the calculation and analysis model into a water pump monitoring system;

[0091] S4012: Real-time traffic estimation via embedded edge computing devices;

[0092] S4013: Use a visualization program written based on the Matplotlib library to display the model calculation results and compare them with the actual measurement values ​​of the flow meter to analyze the errors in real time.

[0093] Specifically, in order to implement flow calculation and analysis based on pump power and head, we built a data-driven computing platform, which includes:

[0094] Data Collection and Modeling: An ABB inverter visual CNC module, pressure sensors, and flowmeters are installed within the pump station's operating environment to collect real-time data on pump power (P), head (H), and pipeline flow (F). The collected data is cleaned and removed of outliers before being constructed into a matrix dataset and stored in a local database. Statistical analysis and multidimensional regression methods are then used to develop a flow calculation model based on this dataset.

[0095] Model deployment and visualization: The computational model is integrated into the water pump monitoring system, and real-time flow estimation is performed through an embedded edge computing device. The model calculation results are displayed using a visualization program written based on the Matplotlib library, and compared with the actual measurement values ​​of the flow meter to analyze errors in real time.

[0096] Error screening and early warning mechanism: During pump operation, if the error between the flow rate estimate of the calculation model and the value measured by the flow meter exceeds the set threshold, the system will trigger an abnormal warning and further screen the data with large errors for analysis to optimize modeling parameters and improve calculation accuracy. This error analysis function can be applied to the pump commissioning stage to help engineers optimize the pump operating conditions and improve commissioning efficiency.

[0097] The above disclosure is merely one or more preferred embodiments of the present application and is not intended to limit the scope of the present application. A person skilled in the art will understand that all or part of the processes of the above embodiments and equivalent changes made in accordance with the claims of the present application are still within the scope of the present application.

Claims

1. A water pump flow simulation calculation method based on nonlinear least squares modeling is characterized by: The following steps are included: Observe and count the power p and lift h of the water pump at the same time, as well as the flow rate f of the corresponding pipeline flowmeter, and construct a data set matrix based on the observed and counted data; Generating a simulation calculation model fitting function based on the data set matrix and observation statistical data; Build a calculation and analysis model based on water pump flow monitoring; The constructed calculation and analysis model is combined with the water pump control flow simulation calculation model to realize flow calculation and analysis based on water pump power and head.

2. The water pump flow simulation calculation method based on nonlinear least squares modeling as claimed in claim 1 is characterized in that: The specific method of forming the data set matrix is: Based on the actual demand statistics of the corresponding data of group i at different times, the pump power p of group i is obtained i , pump head h i And the water pipe flow counter value f i , where: i = {1, 2, ..., n}; The following matrix is ​​constructed based on the obtained n sets of pump power, pump head and pipe flow rate count values: The dataset matrix includes: Water pump dataset matrix: Traffic matrix:

3. The water pump flow simulation calculation method based on nonlinear least squares modeling as claimed in claim 2 is characterized in that: The specific construction method of the simulation calculation model fitting function is: Based on the constructed data set matrix, a binary quadratic nonlinear model is preset, and its function is: f i =c1p i 2 +c2h i 2 +c3p i h i +c4p i +c5h i +c6; The model matrix formula based on the constructed data set matrix combined with the preset binary quadratic nonlinear model is: F=DC, where the parameter matrix C to be determined is: Solving the parameter matrix C in the model matrix formula; Substitute the solved parameter matrix C into the preset binary quadratic nonlinear model to obtain the simulation calculation model fitting function f=c1p 2 +c2h 2 +c3ph+c4p+c5h+c6.

4. The water pump flow simulation calculation method based on nonlinear least squares modeling as claimed in claim 3 is characterized in that: The specific construction method of the calculation and analysis model based on water pump flow monitoring is: Obtain the collected water pump power p and head h, as well as the corresponding pipeline flow meter flow f data; The collected data is cleaned by removing outliers, and then constructed into a matrix data set and stored in the local database; Based on the matrix data set constructed above, a computational analysis model based on water pump flow monitoring is established using statistical analysis and multidimensional regression methods.

5. The water pump flow simulation calculation method based on nonlinear least squares modeling as claimed in claim 4 is characterized in that: The specific method of flow calculation and analysis based on pump power and head is as follows: Exporting the computational analysis model, and visualizing the model appearance and error comparison; Perform data screening and alerts; By screening data with large error values, the accuracy of the simulation calculation model can be improved.

6. The water pump flow simulation calculation method based on nonlinear least squares modeling as claimed in claim 3 is characterized in that: The specific solution method of the parameter matrix C to be determined is: Expand the sum of squared errors: E = ||F-DC|| 2 =(F-DC) T (F-DC), taking the derivative of C at both ends and setting it equal to 0, we get: Simplified to: D T DC=D T F; Based on the above simplified formula D T DC=D T F, solve the parameter matrix C to be determined, if D T D is reversible, so we can solve C: C=(D T D) -1 D T F.

7. The water pump flow simulation calculation method based on nonlinear least squares modeling as claimed in claim 5 is characterized in that: The specific method of visualizing the model appearance and error comparison is as follows: Integrating the computational analysis model into a water pump monitoring system; Real-time traffic estimation through embedded edge computing devices; The model calculation results are displayed using a visualization program written based on the Matplotlib library and compared with the actual measurement values ​​of the flow meter to analyze the errors in real time.

Images