A multi-dimensional analysis method of water quality in thermal network based on big data

Abstract

The present invention discloses a multi-dimensional analysis method for water quality of a thermal network based on big data, and relates to the technical field of water quality analysis of a thermal network. The present invention deploys temperature sensors, flow monitoring equipment and water quality monitors at thermal points to capture dynamic changes in temperature, pressure, flow rate and water quality parameters in real time. A dynamic water quality model adopts a nonlinear autoregressive structure and external input to capture the nonlinear relationship between temperature, pressure, flow rate and historical water quality parameters. Through real-time updating, the model effectively responds to changing environmental conditions and provides accurate water quality predictions. The spatial and temporal distribution of temperature in the network is accurately calculated through temperature field simulation, and is used as input to the dynamic water quality model to more accurately reflect its impact on water quality. The future load change trend is accurately predicted based on historical data and real-time monitoring data through a load prediction model. After the prediction result is combined with the dynamic water quality model, control measures can be taken in advance to prevent sudden water quality problems.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal network water quality analysis, and in particular to a thermal network water quality multi-dimensional analysis method based on big data. Background Art

[0002] In thermal networks, temperature changes have a significant impact on water quality. As the heat in the network continues to circulate, the temperature in the network continues to change, which will directly affect water quality parameters, including pH, dissolved oxygen, and corrosion rate.

[0003] Most current water quality analysis methods assume that water quality parameters are at a constant temperature, while ignoring temperature changes during the thermal cycle, resulting in deviations between water quality analysis results and actual conditions. The operating load of the thermal network also changes dynamically, with load fluctuations varying in different seasons and time periods. Load fluctuations can cause changes in hydraulic conditions within the network, and changes in water flow velocity and pressure can lead to redistribution of dissolved substances, affecting corrosion rates and scaling tendencies within the pipeline. Current water quality management methods mostly regulate water only after water quality problems occur, and are unable to effectively prevent water quality problems.

[0004] Regarding the temperature change problem, most existing technologies estimate water quality changes at different temperatures through regular sampling and laboratory analysis. However, the sampling and analysis process cannot be carried out in real time, and the immediate impact of temperature changes on water quality cannot be reflected in real time. In addition, there are differences in temperature distribution in different areas of the pipe network, making it impossible to perform water quality analysis in different areas. In terms of the operating load changes of the thermal pipe network, most of the current water quality control schemes are mainly based on historical operating data for prediction. According to historical experience, the amount of chemical agents added is adjusted before and after the load increases, or water quality parameters are adjusted when the load changes are expected. However, the lag problem of water quality control measures cannot be improved. Therefore, a multi-dimensional analysis method for thermal pipe network water quality based on big data is urgently needed to solve such problems. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a multi-dimensional analysis method for water quality in a thermal pipeline network based on big data to solve the problems of being unable to respond in time to water quality changes caused by temperature fluctuations, and being unable to effectively identify and process water quality problems in different temperature zones within the pipeline network; lacking real-time analysis and response capabilities for water quality data, and the problem of lag in water quality control measures.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] The present invention provides a multi-dimensional analysis method of water quality in a thermal network based on big data, which includes:

[0009] Step S1: Real-time temperature and load monitoring: deploy temperature sensors and flow monitoring equipment at thermal points to collect pipe network data, and use water quality monitors to monitor parameters of water quality in the pipe network;

[0010] Step S2: monitor the impact of temperature changes, build a dynamic water quality model based on temperature changes, and then use the pipe network data to simulate the temperature field of the thermal pipe network to simulate the water quality changes under different temperature parameters;

[0011] Step S3, monitor the impact of load changes and build a load forecasting model. The model is trained using historical operating data and real-time monitoring data to predict future load change trends. The predicted load data is then combined with a dynamic water quality model to simulate the impact of load changes on water quality.

[0012] Preferably, the heat point refers to the interface location where the heat pipe network is connected to the user;

[0013] The pipe network data includes the water temperature, pressure and flow rate in the pipe network, and the monitored water quality parameter data include pH, dissolved oxygen and corrosion rate;

[0014] Through real-time temperature and load monitoring, dynamic temperature and load changes can be captured and temperature and load dynamic data can be obtained in a timely manner.

[0015] Preferably, the step of constructing a dynamic water quality model in step S2 includes:

[0016] Obtain the sequence data of temperature changing with time, that is, temperature data, temperature data is represented by T(t), pressure data is represented by P(t), flow rate data is represented by v(t), water quality parameter data is represented by W(t), W(t) includes pH value pH(t), dissolved oxygen DO(t), corrosion rate CR(t), t represents time, all data are time series data;

[0017] All data are standardized to eliminate dimensional differences. Here, all data include pipe network data and water quality parameter data. The standardization formula is:

[0018]

[0019] Among them, μ T , μ W , μ P , μ v Respectively represent the means of the corresponding parameters T, W, P and v, σ T , σ W , σ P , σ v is the standard deviation of the corresponding parameters T, W, P and v, T norm (t), W norm (t), P norm(t), v norm (t) represent T, W, P and v after standardization respectively;

[0020] The dynamic water quality model adopts a nonlinear autoregressive model and introduces the NARX model as an external input. The form of the dynamic water quality model is:

[0021] W(t)=f(T norm (t-τ1),…,T norm (t-τ k ),P norm (t-τ1),…,P norm (t-

[0022] τ k ),v norm (t-τ1),…,v norm (t-τ k ),W(t-τ1),…,W(t-τ k ))+∈(t),

[0023] Where W(t) represents the water quality parameters at time t, including pH value pH(t), dissolved oxygen DO(t), and corrosion rate CR(t). f(·) represents the nonlinear function to be estimated, which represents the relationship between temperature, pressure, flow rate and historical water quality parameters. τ1,…,τ k represents the time lag, which is used to capture the time dependence, k represents the time lag order, ∈(t) represents the error term, which is set as white noise;

[0024] The nonlinear function f(·) is a combination of a polynomial and a radial basis function:

[0025]

[0026] Where X represents the vector of input variables, X = [T norm (t-τ1),…,W(t-τ k )],α i , β ij represents the coefficient of the polynomial term, that is, the interaction relationship between the variables, η l 、c l , σ l is the radial basis function RBF parameter, η l represents the weight of each radial basis function, c l represents the center point, σ l represents the width parameter, l represents the corresponding parameter index, i represents the index of the polynomial term, j represents the index of each component in the input variable X, n represents the number of polynomial terms, and m represents the number of input variables.

[0027] Preferably, the step of constructing a dynamic water quality model in step S2 further includes:

[0028] Use gradient descent to minimize the loss function:

[0029] in, is the predicted value of the water quality parameter of the model, λ is the regularization coefficient, T represents the total number of time points of the time series data, n represents the number of polynomial terms, and p represents the number of radial basis functions RBF;

[0030] Collect real-time pipe network data and water quality parameter data, and input them into the dynamic water quality model for real-time correction;

[0031] Update the model parameters using recursive least squares RLS:

[0032] Among them, θ t Contains all model parameters to be estimated, K t represents the gain matrix, which is updated according to the RLS algorithm;

[0033] The dynamic water quality model can not only capture the nonlinear effects of temperature, pressure and flow rate on water quality in the thermal network, but can also be updated in real time to effectively respond to changing environmental conditions.

[0034] Preferably, the step of simulating the temperature field of the thermal network using the network data in step S2 includes:

[0035] Perform geometric modeling and meshing of the pipe network, assuming that the length of each pipe segment is L i , the pipe diameter is D i , where i represents the i-th pipe segment;

[0036] Divide each pipe segment into grid units, and let the length of each unit be Δx, and divide the total Units, x ij =j·Δx,j=1,2,…,N i ,

[0037] where x ij represents the center position of the jth grid cell in the i-th pipe segment;

[0038] Energy conservation analysis is performed on each tiny volume unit in the pipe network, and each tiny volume unit represents a part of the pipe network:

[0039] Where ρ is the fluid density, C p is the specific heat capacity of the fluid, T ij is the temperature of the jth grid cell in the ith pipe segment, Indicates temperature T ij The partial derivative with respect to time t, v iis the flow velocity of the fluid in the ith pipe segment, Indicates temperature T ij The partial derivative with respect to the coordinate x, k is the thermal conductivity of the fluid, q ij is the volume heat source term of the jth grid cell in the ith pipe segment.

[0040] Preferably, the step of simulating the temperature field of the thermal network using the network data in step S2 further includes:

[0041] Set a known temperature T at the inlet in , set the convection boundary condition at the outlet, assuming the external temperature is T ext ,but:

[0042] T i1 =T in ,

[0043] Where h is the convective heat transfer coefficient, T in is the network inlet temperature, T ext is the external ambient temperature, represents the temperature of the grid cell at the end (outlet) of the i-th pipe segment;

[0044] Perform numerical solution and select the time step Δt to satisfy the numerical Courant stability condition:

[0045]

[0046] where v max is the maximum flow velocity in the pipe network;

[0047] Discrete the energy conservation equations by explicit differences to solve the time evolution of the temperature field:

[0048]

[0049] in, is the temperature value at time step n, is the temperature value at the next time step n+1.

[0050] Preferably, the step of simulating the temperature field of the thermal network using the network data in step S2 further includes:

[0051] After solving the temperature distribution of the entire pipe network, each time step As input to dynamic water quality models;

[0052] The temperature field results Input into the dynamic water quality model to analyze the impact of temperature changes on water quality parameters:

[0053]

[0054] Among them, W(t) is the water quality parameter, P(t) is the real-time pressure, and v(t) is the real-time flow rate;

[0055] Simulate the temperature field of the thermal network to provide key input for the dynamic water quality model, capturing the changes in temperature in space and time while also reflecting the impact of such changes on water quality. Based on the simulation results, staff can optimize the network operation parameters and adjust control measures such as chemical agent placement;

[0056] Through the temperature change simulation in step S2, the problem that the traditional static model cannot accurately predict the water quality changes caused by temperature fluctuations is solved. The impact of temperature on water quality is reflected in real time based on the dynamic water quality model, thereby improving the accuracy of water quality prediction.

[0057] Preferably, the historical operation data includes: water flow in the pipe network at different time periods,

[0058] The temperature of different areas in the pipe network in different seasons and weather conditions.

[0059] Pressure data of each thermal point, including pressure fluctuation and change trend,

[0060] and energy input and output data.

[0061] Preferably, the real-time monitoring data, i.e., the data in real-time temperature and load monitoring, includes pipe network data and water quality parameters.

[0062] Preferably, the load forecasting model is constructed in step S3 as follows:

[0063] Collect historical operation data and perform cleaning, normalization and feature extraction on it;

[0064] Select ARIMA, LSTM and random forest regression models and train them based on historical data;

[0065] When training the model, the load, i.e., the parameters in the pipe network data and water quality parameter data, are used as target variables, and the input variables include time, temperature, and seasonal characteristics;

[0066] After evaluating the model performance using the cross-validation method, the load forecasting model is used to perform load forecasting;

[0067] The predicted future load data is input into the dynamic water quality model. The dynamic water quality model is used to simulate the impact of load changes on water quality parameters. Based on the prediction results of the load prediction model, the water quality control strategy is determined in advance, and the water quality is actively adjusted before the load changes to prevent sudden water quality problems.

[0068] The beneficial effects of the present invention are:

[0069] The present invention deploys temperature sensors, flow monitoring equipment and water quality monitors at thermal points to capture the dynamic changes of temperature, pressure, flow rate and water quality parameters in real time. The dynamic water quality model adopts a nonlinear autoregressive structure and external input to capture the nonlinear relationship between temperature, pressure, flow rate and historical water quality parameters. Through real-time updating, the model effectively responds to changing environmental conditions, thereby providing accurate water quality predictions. The temperature field simulation is used to accurately calculate the spatial and temporal distribution of temperature in the pipe network, which is used as the input of the dynamic water quality model to more accurately reflect its impact on water quality. The load prediction model is used to accurately predict future load change trends based on historical data and real-time monitoring data. After the prediction results are combined with the dynamic water quality model, regulatory measures can be taken in advance to prevent sudden water quality problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0071] Figure 1 It is a schematic flow chart of the multi-dimensional analysis method of water quality in thermal pipe networks of the present invention. DETAILED DESCRIPTION

[0072] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0073] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0074] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0075] Example 1, reference Figure 1 This embodiment provides a multi-dimensional analysis method for water quality in a thermal network based on big data, comprising the following steps:

[0076] Step S1: Real-time temperature and load monitoring: deploy temperature sensors and flow monitoring equipment at thermal points to collect pipe network data, and use water quality monitors to monitor parameters of water quality in the pipe network;

[0077] The heating point refers to the interface location where the heating network connects to the user;

[0078] The pipe network data includes the water temperature, pressure and flow rate in the pipe network, and the monitored water quality parameter data include pH, dissolved oxygen and corrosion rate;

[0079] Through real-time temperature and load monitoring, dynamic temperature and load changes can be captured, and temperature and load dynamic data can be obtained in a timely manner;

[0080] Specifically, through real-time temperature and load monitoring, dynamic capture of key parameters in the pipeline network can be achieved. In traditional methods, water quality analysis often relies on regular sampling and laboratory analysis, which has a large lag and is difficult to reflect changes in the pipeline network in real time. This method deploys temperature sensors and flow monitoring equipment at thermal points, combined with water quality monitors, to achieve real-time monitoring of temperature, pressure, flow rate and water quality parameters (such as pH, dissolved oxygen, and corrosion rate).

[0081] Step S2: monitor the impact of temperature changes, build a dynamic water quality model based on temperature changes, and then use the pipe network data to simulate the temperature field of the thermal pipe network to simulate the water quality changes under different temperature parameters;

[0082] The steps of constructing a dynamic water quality model in step S2 include:

[0083] Obtain the sequence data of temperature changing with time, that is, temperature data, temperature data is represented by T(t), pressure data is represented by P(t), flow rate data is represented by v(t), water quality parameter data is represented by W(t), W(t) includes pH value pH(t), dissolved oxygen DO(t), corrosion rate CR(t), t represents time, all data are time series data;

[0084] All data are standardized to eliminate dimensional differences. All data here include pipe network data and water quality parameter data. The standardization formula is:

[0085]

[0086] Among them, μ T , μ w , μ P , μ v Respectively represent the means of the corresponding parameters T, W, P and v, σ T , σ W , σ p , σ v is the standard deviation of the corresponding parameters T, W, P and v, Tnorm (t), W norm (t), P norm (t), v norm (t) represent T, W, P and v after standardization respectively;

[0087] The dynamic water quality model adopts a nonlinear autoregressive model and introduces the NARX model as an external input. The form of the dynamic water quality model is:

[0088] W(t)=f(T norm (t-τ1),…,T norm (t-τ k ),P norm (t-τ1),…,P norm (t-τ k ),v norm (t-τ1),…,v norm (t-τ k ),W(t-τ1),…,W(t-τ k ))+∈(t),

[0089] Where W(t) represents the water quality parameters at time t, including pH value pH(t), dissolved oxygen DO(t), and corrosion rate CR(t). f(·) represents the nonlinear function to be estimated, which represents the relationship between temperature, pressure, flow rate and historical water quality parameters. τ1,…,τ k represents the time lag, which is used to capture the time dependence, k represents the time lag order, ∈(t) represents the error term, which is set as white noise;

[0090] The nonlinear function f(·) is a combination of a polynomial and a radial basis function:

[0091]

[0092] Where X represents the vector of input variables, X = [T norm (t-τ1),…,W(t-τ k )],α i , β ij represents the coefficient of the polynomial term, that is, the interaction relationship between the variables, η l 、c l , σ l is the radial basis function RBF parameter, η l represents the weight of each radial basis function, c l represents the center point, σ l represents the width parameter, l represents the corresponding parameter index, i represents the index of the polynomial term, j represents the index of each component in the input variable X, n represents the number of polynomial terms, and m represents the number of input variables;

[0093] The step of constructing a dynamic water quality model in step S2 also includes:

[0094] Use gradient descent to minimize the loss function:

[0095] in, is the predicted value of the water quality parameter of the model, λ is the regularization coefficient, T represents the total number of time points of the time series data, n represents the number of polynomial terms, and p represents the number of radial basis functions RBF;

[0096] Collect real-time pipe network data and water quality parameter data, and input them into the dynamic water quality model for real-time correction;

[0097] Update the model parameters using recursive least squares RLS:

[0098] Among them, θ t Contains all model parameters to be estimated, K t represents the gain matrix, which is updated according to the RLS algorithm;

[0099] The dynamic water quality model can not only capture the nonlinear effects of temperature, pressure, and flow rate on water quality in the thermal network, but also be updated in real time to effectively respond to changing environmental conditions;

[0100] Specifically, the dynamic water quality model based on temperature changes greatly improves the accuracy of water quality predictions. The model adopts a nonlinear autoregressive structure and introduces an external input NARX model, which can capture the nonlinear relationship between temperature, pressure, flow rate and historical water quality parameters; through temperature field simulation, it predicts its immediate impact on water quality. The dynamic modeling method effectively solves the problem that traditional static models cannot accurately predict water quality changes caused by temperature fluctuations, making water quality predictions closer to actual operating conditions.

[0101] The step of simulating the temperature field of the thermal network using the network data in step S2 includes:

[0102] Perform geometric modeling and meshing of the pipe network, assuming that the length of each pipe segment is L i , the pipe diameter is D i , where i represents the i-th pipe segment;

[0103] Divide each pipe segment into grid units, and let the length of each unit be Δx, and divide the total units,

[0104] x ij =j·Δx,j=1,2,…,N i ,

[0105] where x ijrepresents the center position of the jth grid cell in the i-th pipe segment;

[0106] Energy conservation analysis is performed on each tiny volume unit in the pipe network, and each tiny volume unit represents a part of the pipe network:

[0107]

[0108] Where ρ is the fluid density, C p is the specific heat capacity of the fluid, T ij is the temperature of the jth grid cell in the ith pipe segment, Indicates temperature T ij The partial derivative with respect to time t, v i is the flow velocity of the fluid in the ith pipe segment, Indicates temperature T ij The partial derivative with respect to the coordinate x, k is the thermal conductivity of the fluid, q ij is the volume heat source term of the jth grid cell in the i-th pipe segment;

[0109] The step of simulating the temperature field of the thermal network using the network data in step S2 also includes:

[0110] Set a known temperature T at the inlet in , set the convection boundary condition at the outlet, assuming the external temperature is T ext ,but:

[0111] T i1 =T in ,

[0112] Where h is the convective heat transfer coefficient, T in is the network inlet temperature, T ext is the external ambient temperature, represents the temperature of the grid cell at the end (outlet) of the i-th pipe segment;

[0113] Perform numerical solution and select the time step Δt to satisfy the numerical Courant stability condition:

[0114]

[0115] where v max is the maximum flow velocity in the pipe network;

[0116] Discrete the energy conservation equations by explicit differences to solve the time evolution of the temperature field:

[0117]

[0118] in, is the temperature value at time step n, is the temperature value at the next time step n+1;

[0119] Specifically, the thermal network is geometrically modeled and meshed, and the energy conservation equation is numerically solved to simulate the temperature distribution inside the network. The simulation results are then input into the dynamic water quality model, allowing the model to more accurately reflect the impact of temperature changes on water quality and improve the accuracy of the prediction.

[0120] The step of simulating the temperature field of the thermal network using the network data in step S2 also includes:

[0121] After solving the temperature distribution of the entire pipe network, each time step As input to dynamic water quality models;

[0122] The temperature field results Input into the dynamic water quality model to analyze the impact of temperature changes on water quality parameters:

[0123]

[0124] Among them, W(t) is the water quality parameter, P(t) is the real-time pressure, and v(t) is the real-time flow rate;

[0125] Simulate the temperature field of the thermal network to provide key input for the dynamic water quality model, capturing the changes in temperature in space and time while also reflecting the impact of such changes on water quality. Based on the simulation results, staff can optimize the network operation parameters and adjust control measures such as chemical agent placement;

[0126] Through the temperature change simulation in step S2, the problem that the traditional static model cannot accurately predict the water quality changes caused by temperature fluctuations is solved. The influence of temperature on water quality is reflected in real time based on the dynamic water quality model, which improves the accuracy of water quality prediction.

[0127] Step S3: monitor the impact of load changes and build a load forecasting model. The model is trained using historical operation data and real-time monitoring data to predict future load change trends. The predicted load data is then combined with a dynamic water quality model to simulate the impact of load changes on water quality.

[0128] Historical operation data includes: water flow in the pipe network at different time periods,

[0129] The temperature of different areas in the pipe network in different seasons and weather conditions.

[0130] Pressure data of each thermal point, including pressure fluctuation and change trend,

[0131] and energy input and output data;

[0132] Real-time monitoring data refers to the data from real-time temperature and load monitoring, including pipe network data and water quality parameters;

[0133] The method of constructing the load forecasting model in step S3 is:

[0134] Collect historical operation data and perform cleaning, normalization and feature extraction on it;

[0135] Select ARIMA, LSTM and random forest regression models and train them based on historical data;

[0136] When training the model, the load, i.e., the parameters in the pipe network data and water quality parameter data, are used as target variables, and the input variables include time, temperature, and seasonal characteristics;

[0137] After evaluating the model performance using the cross-validation method, the load forecasting model is used to perform load forecasting;

[0138] The predicted future load data is input into the dynamic water quality model, and the impact of load changes on water quality parameters is simulated through the dynamic water quality model;

[0139] Specifically, the load forecasting model constructed uses historical operation data and real-time monitoring data to predict future load change trends through time series forecasting and machine learning algorithms, combines the predicted load data with the dynamic water quality model, simulates the impact of load changes on water quality, and further optimizes water quality management. Based on the prediction results of the load forecasting model, the water quality control strategy is determined in advance, and the water quality is actively adjusted before the load changes to prevent sudden water quality problems.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A multi-dimensional analysis method for water quality in a thermal network based on big data, characterized by: include, Step S1: Real-time temperature and load monitoring: deploy temperature sensors and flow monitoring equipment at thermal points to collect pipe network data, and use water quality monitors to monitor parameters of water quality in the pipe network; Step S2: monitor the impact of temperature changes, build a dynamic water quality model based on temperature changes, and then use the pipe network data to simulate the temperature field of the thermal pipe network to simulate the water quality changes under different temperature parameters; Step S3: monitor the impact of load changes and build a load forecasting model. The model is trained using historical operation data and real-time monitoring data to predict future load change trends. The predicted load data is then combined with a dynamic water quality model to simulate the impact of load changes on water quality. The steps of constructing a dynamic water quality model in step S2 include: Obtain the sequence data of temperature changing with time, that is, temperature data, temperature data is represented by T(t), pressure data is represented by P(t), flow rate data is represented by v(t), water quality parameter data is represented by W(t), W(t) includes pH value pH(t), dissolved oxygen DO(t), corrosion rate CR(t), t represents time, all data are time series data; All data are standardized to eliminate dimensional differences. Here, all data include pipe network data and water quality parameter data. The standardization formula is: Among them, μ T , μ W , μ P , μ v Respectively represent the means of the corresponding parameters T(t), W(t), P(t) and v(t), σ T , σ W , σ P , σ v is the standard deviation of the corresponding parameters T(t), W(t), P(t) and v(t), T norm (t), W norm (t), P norm (t), v norm (t) represent T(t), W(t), P(t) and v(t) after standardization respectively; The dynamic water quality model adopts a nonlinear autoregressive model and introduces the NARX model as an external input. The form of the dynamic water quality model is: W(t)=f(T norm (t-τ1),…,T norm (t-τ k ),P norm (t-τ1),…,P norm (t-τ k ),v norm (t-τ1),…,v norm (t-τ k ),W(t-τ1),…,W(t-τ k ))+∈(t), Where W(t) represents the water quality parameters at time t, including pH value pH(t), dissolved oxygen DO(t), and corrosion rate CR(t). f(·) represents the nonlinear function to be estimated, which represents the relationship between temperature, pressure, flow rate and historical water quality parameters. τ1,…,τ k represents the time lag, which is used to capture the time dependence, k represents the time lag order, ∈(t) represents the error term, which is set as white noise; The nonlinear function f(·) is a combination of a polynomial and a radial basis function: Where X represents the vector of input variables, X = [T norm (t-τ1),…,W(t-τ k )],α i′ , β i′j′ represents the coefficient of the polynomial term, that is, the interaction relationship between the variables, η l 、c l , σ l is the radial basis function RBF parameter, η l represents the weight of each radial basis function, c l represents the center point, σ l represents the width parameter, l represents the corresponding parameter index, i′ represents the index of the polynomial term, j′ represents the index of each component in the input variable X, n represents the number of polynomial terms, m represents the number of input variables, and p represents the number of radial basis functions RBF.

2. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 1, characterized in that: The heating point refers to the interface location where the heating network connects to the user; The pipe network data includes the water temperature, pressure and flow rate in the pipe network, and the monitored water quality parameter data include pH, dissolved oxygen and corrosion rate.

3. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 2, characterized in that: The step of constructing a dynamic water quality model in step S2 also includes: Use gradient descent to minimize the loss function: in, is the predicted value of the water quality parameter of the model, λ is the regularization coefficient, T represents the total number of time points of the time series data, and n represents the number of polynomial terms; Collect real-time pipe network data and water quality parameter data, and input them into the dynamic water quality model for real-time correction; Update the model parameters using recursive least squares RLS: Among them, θ t Contains all parameters of the dynamic water quality model to be estimated, K t Represents the gain matrix, which is updated according to the RLS algorithm.

4. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 3, characterized in that: The step of simulating the temperature field of the thermal network using the network data in step S2 includes: Perform geometric modeling and meshing of the pipe network, assuming that the length of each pipe segment is L i , the pipe diameter is D i , where i represents the i-th pipe segment; Divide each pipe segment into grid units, and let the length of each unit be Δx, and divide the total Units, x ij =j·Δx,j=1,2,…,N i , where x ij represents the center position of the jth grid cell in the i-th pipe segment; Energy conservation analysis is performed on each tiny volume unit in the pipe network, and each tiny volume unit represents a part of the pipe network: Where ρ is the fluid density, C p is the specific heat capacity of the fluid, T ij is the temperature of the jth grid cell in the ith pipe segment, Indicates temperature T ij The partial derivative with respect to time t, v i is the flow velocity of the fluid in the ith pipe segment, Indicates temperature T ij The partial derivative with respect to the coordinate x, k is the thermal conductivity of the fluid, q ij is the volume heat source term of the jth grid cell in the ith pipe segment.

5. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 4, characterized in that: The step of simulating the temperature field of the thermal network using the network data in step S2 also includes: Set a known temperature T at the inlet in , set the convection boundary condition at the outlet, assuming the external temperature is T ext ,but: T i1 =T in , Where h is the convective heat transfer coefficient, T in is the network inlet temperature, T ext is the external ambient temperature, T iNi represents the temperature of the last grid cell in the i-th pipe segment; Perform numerical solution and select the time step Δt to satisfy the numerical Courant stability condition: where v max is the maximum flow velocity in the pipe network; Discrete the energy conservation equations by explicit differences to solve the time evolution of the temperature field: in, is the temperature value at time step n, is the temperature value at the next time step n+1.

6. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 5, characterized in that: The step of simulating the temperature field of the thermal network using the network data in step S2 also includes: After solving the temperature distribution of the entire pipe network, each time step As input to dynamic water quality models; The temperature field results Input into the dynamic water quality model to analyze the impact of temperature changes on water quality parameters: Among them, W(t) is the water quality parameter, P(t) is the real-time pressure, and v(t) is the real-time flow rate.

7. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 6, characterized in that: Historical operation data includes: water flow in the pipe network at different time periods, The temperature of different areas in the pipe network in different seasons and weather conditions. Pressure data of each thermal point, including pressure fluctuation and change trend, and energy input and output data.

8. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 7, characterized in that: Real-time monitoring data refers to the data from real-time temperature and load monitoring, including pipe network data and water quality parameters.

9. A multi-dimensional analysis method for water quality in a thermal network based on big data as claimed in claim 8, characterized in that: The method of constructing the load forecasting model in step S3 is: Collect historical operation data and perform cleaning, normalization and feature extraction on it; Select ARIMA, LSTM and random forest regression models and train them based on historical data; When training the model, the load, i.e., the parameters in the pipe network data and water quality parameter data, are used as target variables, and the input variables include time, temperature, and seasonal characteristics; After evaluating the model performance using the cross-validation method, the load forecasting model is used to perform load forecasting; The predicted future load data is input into the dynamic water quality model, and the impact of load changes on water quality parameters is simulated through the dynamic water quality model.

Images