CN117196159A - Intelligent water service partition metering system based on Internet big data analysis - Google Patents

Abstract

The invention relates to the technical field of resource management systems, in particular to an intelligent water service partition metering system based on internet big data analysis. According to the invention, the water affair data collected in real time through the Internet of things technology and the sensor equipment is more accurate and timely, a plurality of algorithms and models such as a neural network, a random forest and a time sequence are adopted, more accurate water pressure adjustment, water quality prediction and water resource trend analysis can be performed, the simulation environment is more real by combining the digital twin technology, more effective test and optimization of the strategies by a manager are allowed, the comprehensive water affair analysis of the system from a plurality of angles is also ensured by the fusion of the multi-source data, and the decision is more comprehensive and scientific.

Description

Smart water district metering system based on Internet big data analysis

Technical field

The present invention relates to the technical field of resource management systems, and in particular to a smart water district metering system based on Internet big data analysis.

Background technique

The main focus of the resource management system is how to manage and configure various resources more effectively. In many industries, especially for the management of necessary resources (such as water, electricity, gas, etc.), resource management systems can help achieve optimal utilization of resources, reduce waste, and improve efficiency. This technical field includes all aspects of resource collection, analysis, distribution, and monitoring.

The smart water district metering system based on Internet big data analysis is a water management system that integrates big data analysis and Internet technology. It can collect, process and analyze data sent from various water divisions in real time, such as water flow, water quality, equipment operating status, etc. By analyzing this data, the system can provide operators with in-depth insights and decision support regarding water resources management. The main purpose is to achieve intelligent management of water affairs. This can not only improve the efficiency of water use and reduce water waste, but also help predict and respond to various water-related problems, such as water quality decline, equipment failure, leakage, etc.

Existing systems have some obvious shortcomings in handling complex water scenarios. First, existing systems often rely on manually or regularly collected data, which leads to data delays and discontinuities, which in turn affects the timeliness of management decisions. Secondly, most existing systems do not integrate multi-source data, such as satellite and meteorological data, resulting in a lack of multi-dimensional consideration in analysis and prediction, causing some important factors to be ignored. Furthermore, the simulation and emulation technology of existing systems is often not mature enough, leading to deviations in practical applications. In addition, detection and response to abnormal events are not fast and accurate enough, resulting in delayed handling of leaks or other problems.

Contents of the invention

The purpose of this invention is to solve the shortcomings in the existing technology and propose a smart water district metering system based on Internet big data analysis.

In order to achieve the above purpose, the present invention adopts the following technical solution: a smart water district metering system based on Internet big data analysis includes a data collection module, a water pressure adjustment module, a water quality monitoring module, a multi-source data fusion module, a data prediction module, and a virtual simulation module, anomaly detection module, notification feedback module;

The data collection module is based on Internet of Things technology and uses sensor equipment to collect real-time water affairs data, including water pressure, flow and water quality information, and generate real-time water affairs data summary;

The water pressure adjustment module is based on real-time water affairs data summary, uses a neural network model to analyze water pressure leakage and consumption patterns, and performs adaptive water pressure adjustment to generate optimized water pressure parameters;

The water quality monitoring module is based on real-time water affairs data summary, uses random forest algorithm to predict water quality changes, and uses GIS for data visualization to generate water quality prediction reports;

The multi-source data fusion module combines real-time water affairs data summary with satellite and meteorological data, uses deep learning technology to perform data fusion, conducts water affairs analysis, and generates a comprehensive water affairs analysis report;

The data prediction module is based on the comprehensive water analysis report, uses a time series model to predict water resources conditions, conducts trend analysis, and generates a water resources prediction report;

The virtual simulation module uses digital twin technology to simulate water networks based on water resource prediction reports, conducts management strategy testing, and generates simulation test reports;

The anomaly detection module is based on the comprehensive water analysis report, uses a machine learning model to identify abnormal consumption and leakage patterns, and generates an abnormal event report;

The notification feedback module collects maintenance team notifications and problem feedback based on abnormal event reports and simulation test reports, and generates maintenance feedback records;

The real-time water data summary is specifically multi-node water pressure, flow and water quality data stored in the form of time series, including temperature, pH value, and turbidity. The optimized water pressure parameters are specifically water pumps adjusted based on consumption patterns and leakage conditions. Operating parameters, the water quality prediction report specifically predicts the water quality change trend of multiple nodes in the future time period, the comprehensive water analysis report includes pump station operation status, water pipe network structure, water quality status, user consumption pattern, the water resource prediction The report is specifically the predicted changes and demand trends of water resources in the future time period, the simulation test report is specifically the test results and optimization plans, and the abnormal event report is specifically the real-time pipe network leakage data and abnormal consumption patterns.

As a further solution of the present invention: the data collection module includes a pressure sensing sub-module, a flow sensing sub-module, and a water quality sensing sub-module;

The water pressure adjustment module includes a time series data analysis sub-module, a leak detection sub-module, and a pressure adaptive adjustment sub-module;

The water quality monitoring module includes a multi-dimensional data analysis sub-module, a water quality parameter prediction sub-module, and a GIS visualization sub-module;

The multi-source data fusion module includes a remote sensing image analysis sub-module, a meteorological data processing sub-module, and a fusion algorithm application sub-module;

The data prediction module includes a short-term prediction sub-module, a long-term prediction sub-module, and a trend analysis sub-module;

The virtual simulation module includes a digital model establishment sub-module, a strategy testing sub-module, and an optimization plan sub-module;

The anomaly detection module includes a leak identification sub-module, a consumption pattern analysis sub-module, and a real-time alarm sub-module;

The notification feedback module includes an alarm notification sub-module, a team response sub-module, and a maintenance feedback integration sub-module.

As a further solution of the present invention: the pressure sensing sub-module is based on Internet of Things technology and uses a differential pressure detection algorithm to monitor pressure changes in the water in real time, perform data analysis, and generate a real-time water pressure data report;

The flow sensing sub-module is based on the real-time water pressure data report, uses the turbine flow calculation method to monitor the flow rate and flow rate of the water flow, and combines the water pressure data with the flow estimation to generate a real-time flow data report;

The water quality sensing sub-module is based on real-time flow data reporting and uses spectral detection methods to analyze the chemical components in the water, conduct water quality assessment, and generate real-time water quality data reports;

The differential pressure detection algorithm specifically performs differential operations on water pressure data continuously to obtain the changing trend of water pressure. The real-time water pressure data report includes pressure values, pressure changing trends and abnormal fluctuations. The turbine flow calculation method is specifically: The flow value is calculated according to the rotation speed of the turbine. The real-time flow data report includes flow speed, flow value and flow change trend. The spectral detection method is specifically to analyze the spectral characteristics of the water sample through a spectroscopic instrument. The real-time water quality data report specifically refers to Evaluation report of chemical composition, turbidity, pH value.

As a further solution of the present invention: the time series data analysis sub-module is based on real-time water affairs data summary and uses a time series prediction algorithm to predict future water pressure changes, formulate adjustment strategies, and generate a water pressure change prediction report;

The leakage detection sub-module is based on the water pressure change prediction report and uses an anomaly detection algorithm to analyze abnormal drops in the water pressure data, determine the leakage risk, formulate a repair plan, and generate a leakage detection report;

The pressure adaptive adjustment sub-module is based on the leak detection report and uses fuzzy logic control method to automatically adjust the water pressure according to the leakage risk and consumption pattern, and generate optimized water pressure parameters;

The time series prediction algorithm specifically uses an ARIMA or LSTM model to model and predict historical water pressure data. The water pressure change prediction report includes the predicted water pressure value, prediction accuracy and adjustment strategy. The anomaly detection algorithm Specifically, the One-Class SVM or Isolation Forest method is used to identify abnormal points. The leakage detection report includes the abnormal location, leakage degree and emergency repair plan. The fuzzy logic control method specifically makes decisions based on fuzzy sets and fuzzy rules.

As a further solution of the present invention: the multi-dimensional data analysis sub-module is based on real-time water affairs data summary, uses K-means clustering algorithm to perform dimensional analysis of water quality data, and mines high-frequency problem areas to generate water quality problem clustering results;

The water quality parameter prediction sub-module uses the random forest algorithm to predict the trend of water quality parameters based on the clustering results of water quality problems, and generates water quality parameter prediction results;

The GIS visualization sub-module uses GIS technology to visually present geographical information based on the water quality parameter prediction results, and generates a water quality prediction report;

The K-means clustering algorithm specifically classifies water quality data into multi-level quality categories. The random forest algorithm includes ensemble learning of decision trees. The water quality parameter prediction results are specifically the water quality changes in the future time period. Trend, the GIS technology is specifically geographical information system technology.

As a further solution of the present invention: the remote sensing image analysis sub-module uses a convolutional neural network to extract remote sensing image features based on remote sensing image data acquired by satellites, and generates a remote sensing image feature report;

The meteorological data processing sub-module uses the data preprocessing method to clean the data based on the remote sensing image feature report and meteorological data, and generates a processed meteorological data report;

The fusion algorithm application sub-module is based on the processed meteorological data report and real-time water affairs data summary, and uses deep learning technology to perform fusion analysis of multi-source data to generate a comprehensive water affairs analysis report;

The convolutional neural network is specifically a feedforward neural network, used for image and sound recognition. The remote sensing image feature report includes feature values and feature vectors extracted from the image. The data preprocessing method includes missing value processing, abnormality Value detection and data standardization, the deep learning technology is specifically a multi-layer neural network model, used to process non-linear relationships.

As a further solution of the present invention: the short-term prediction sub-module is based on the comprehensive water analysis report and uses the autoregressive integral moving average model to perform short-term predictions on recent water resources data and generate short-term water resources prediction data;

The long-term prediction sub-module is based on short-term water resources prediction data and uses a long short-term memory network to predict medium- and long-term water resources conditions and generate long-term water resources prediction data;

The trend analysis sub-module uses linear regression analysis to conduct water resource development trends and potential risk assessments based on long-term water resources prediction data, and generates a water resources prediction report;

The autoregressive integrated moving average model is used to capture the autoregressive and moving average characteristics of data, and the linear regression analysis specifically uses mathematical methods to model and analyze the linear relationship between variables.

As a further solution of the present invention: the digital model establishment sub-module uses three-dimensional modeling technology to establish a digital water network model based on the water resource prediction report and generate a digital water network model;

The strategy testing sub-module is based on the digital water network model and uses the Monte Carlo simulation method to test the effect of the water resources management strategy and generate strategy test results;

The optimization plan sub-module uses the decision tree analysis method to propose the optimal management strategy and optimization plan based on the strategy test results, and generates a simulation test report;

The three-dimensional modeling technology specifically refers to the use of computer-aided design software to create geometric representations of spatial objects. The Monte Carlo simulation method simulates and analyzes future results by randomly extracting parameters from probability distributions. The decision tree analysis method specifically As a decision support tool, use tree diagrams and forecast results to analyze probabilistic event outcomes, resource costs and benefits.

As a further solution of the present invention: based on the comprehensive water analysis report, the leakage identification sub-module uses a convolutional neural network to analyze water pipeline image data, identify leakage characteristics, and generate a leakage identification report;

The consumption pattern analysis sub-module uses the K-means clustering algorithm to analyze consumer water usage data based on leakage identification reports, detect abnormal consumption patterns, and generate abnormal consumption pattern reports;

The real-time alarm sub-module is based on abnormal consumption pattern reporting and uses the threshold analysis method for real-time data monitoring. When the data exceeds the preset threshold, an alarm is triggered and an abnormal event report is generated;

The convolutional neural network specifically uses deep learning technology to extract features from pipeline system images and identify leak locations and sizes. The K-means clustering algorithm includes clustering user consumption data based on multi-dimensional data including water consumption and time. The data divides user groups and identifies abnormal consumption patterns. The threshold analysis method specifically refers to setting the safe range of key indicators of water flow and pressure. When the real-time data exceeds the safe range, an alarm is automatically triggered.

As a further solution of the present invention: the alarm notification sub-module uses automatic message push technology to notify the maintenance team of alarms based on abnormal event reports and simulation test reports, and generate alarm notification records;

The team response sub-module uses instant messaging software to coordinate within the team based on alarm notification records, formulate emergency response plans, and generate team response records;

The maintenance feedback integration sub-module is based on team response records, uses data fusion technology to integrate maintenance feedback, including problem solving progress and effects, and generates maintenance feedback records.

Compared with the existing technology, the advantages and positive effects of the present invention are:

In the present invention, water affairs data collected in real time through Internet of Things technology and sensor equipment makes data acquisition more accurate and timely. Using various algorithms and models such as neural networks, random forests and time series, more accurate water pressure adjustment, water quality prediction and water resource trend analysis can be carried out. Combined with digital twin technology, the simulation environment is more realistic, allowing managers to test and optimize strategies more effectively. The fusion of multi-source data also ensures that the system conducts comprehensive water affairs analysis from multiple perspectives, making decision-making more comprehensive and scientific.

Description of the drawings

Figure 1 is a system flow chart of the present invention;

Figure 2 is a schematic diagram of the system framework of the present invention;

Figure 3 is a flow chart of the data collection module of the present invention;

Figure 4 is a flow chart of the water pressure adjustment module of the present invention;

Figure 5 is a flow chart of the water quality monitoring module of the present invention;

Figure 6 is a flow chart of the multi-source data fusion module of the present invention;

Figure 7 is a flow chart of the data prediction module of the present invention;

Figure 8 is a flow chart of the virtual simulation module of the present invention;

Figure 9 is a flow chart of the anomaly detection module of the present invention;

Figure 10 is a flow chart of the notification feedback module of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention and are not intended to limit the present invention.

In the description of the present invention, it should be understood that the terms "length", "width", "upper", "lower", "front", "back", "left", "right", "vertical", The orientations or positional relationships indicated by "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description. It is not intended to indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore is not to be construed as a limitation of the invention. In addition, in the description of the present invention, "plurality" means two or more than two, unless otherwise clearly and specifically limited.

Embodiment 1: Please refer to Figure 1. A smart water district metering system based on Internet big data analysis includes a data collection module, a water pressure adjustment module, a water quality monitoring module, a multi-source data fusion module, a data prediction module, a virtual simulation module, and anomaly detection. Module, notification feedback module;

The data collection module is based on Internet of Things technology and uses sensor equipment to collect real-time water affairs data, including water pressure, flow and water quality information, and generate real-time water affairs data summary;

The water pressure adjustment module is based on real-time water affairs data collection, uses a neural network model to analyze water pressure leakage and consumption patterns, and performs adaptive water pressure adjustments to generate optimized water pressure parameters;

The water quality monitoring module is based on real-time water affairs data summary, uses random forest algorithm to predict water quality changes, and uses GIS for data visualization to generate water quality prediction reports;

The multi-source data fusion module combines real-time water affairs data summary with satellite and meteorological data, uses deep learning technology to perform data fusion, conducts water affairs analysis, and generates a comprehensive water affairs analysis report;

The data prediction module is based on the comprehensive water analysis report, uses a time series model to predict water resources conditions, conducts trend analysis, and generates a water resources prediction report;

The virtual simulation module uses digital twin technology to simulate water networks based on water resource forecast reports, conduct management strategy testing, and generate simulation test reports;

The anomaly detection module is based on the comprehensive water analysis report, uses machine learning models to identify abnormal consumption and leakage patterns, and generates abnormal event reports;

The notification feedback module collects maintenance team notifications and problem feedback based on abnormal event reports and simulation test reports, and generates maintenance feedback records;

Real-time water affairs data summary is specifically multi-node water pressure, flow and water quality data stored in the form of time series, including temperature, pH value, turbidity. Optimized water pressure parameters are specifically water pump operating parameters adjusted based on consumption patterns and leakage conditions. Water quality The prediction report specifically predicts the water quality change trend of multiple nodes in the future time period. The comprehensive water analysis report includes the operating status of pump stations, water pipe network structure, water quality conditions, and user consumption patterns. The water resource prediction report specifically predicts the water quality change trend in the future time period. Resource changes and demand trends, simulation test reports specifically include test results and optimization plans, and abnormal event reports include real-time pipe network leakage data and abnormal consumption patterns.

Through real-time data collection, intelligent analysis and simulation testing, the system can realize real-time monitoring, intelligent adjustment and optimal management of the water supply system. The water pressure adjustment module can adaptively adjust water pressure parameters to improve the efficiency and stability of the water supply system; the water quality monitoring module can predict water quality change trends, promptly discover water quality problems and take corresponding measures to improve them; the multi-source data fusion module can comprehensively understand the water supply The operating status and trends of the system provide scientific basis for decision-making; the data prediction module can predict water resources conditions and management strategy effects in advance, helping decision-makers make scientific plans; the virtual simulation module can evaluate the effects of different management strategies and optimize the water supply system Operation efficiency and resource utilization; the anomaly detection module can timely detect pipe network leaks and abnormal consumption, reducing resource waste and loss; the notification feedback module can quickly respond and solve problems in the water supply system, improving service quality and user satisfaction.

Please refer to Figure 2. The data collection module includes a pressure sensing sub-module, a flow sensing sub-module, and a water quality sensing sub-module;

The water pressure adjustment module includes a time series data analysis sub-module, a leak detection sub-module, and a pressure adaptive adjustment sub-module;

The water quality monitoring module includes multi-dimensional data analysis sub-module, water quality parameter prediction sub-module, and GIS visualization sub-module;

The multi-source data fusion module includes a remote sensing image analysis sub-module, a meteorological data processing sub-module, and a fusion algorithm application sub-module;

The data prediction module includes short-term prediction sub-module, long-term prediction sub-module and trend analysis sub-module;

The virtual simulation module includes a digital model establishment sub-module, a strategy testing sub-module, and an optimization plan sub-module;

The anomaly detection module includes a leak identification sub-module, a consumption pattern analysis sub-module, and a real-time alarm sub-module;

The notification feedback module includes an alarm notification sub-module, a team response sub-module, and a maintenance feedback integration sub-module.

In the data collection module, the pressure sensing submodule is responsible for collecting water pressure data in real time, the flow sensing submodule is responsible for collecting flow data in real time, and the water quality sensing submodule is responsible for collecting water quality information in real time.

In the water pressure adjustment module, the time series data analysis sub-module performs time series analysis on real-time water affairs data, the leakage detection sub-module uses the neural network model to perform water pressure leakage analysis, and the pressure adaptive adjustment sub-module performs adaptive adjustment of water pressure based on the analysis results. .

In the water quality monitoring module, the multi-dimensional data analysis sub-module conducts multi-dimensional analysis of real-time water affairs data, the water quality parameter prediction sub-module uses the random forest algorithm to predict water quality changes, and the GIS visualization sub-module displays the data visually.

In the multi-source data fusion module, the remote sensing image analysis sub-module analyzes satellite data, the meteorological data processing sub-module processes meteorological data, and the fusion algorithm application sub-module fuses real-time water affairs data with satellite and meteorological data.

In the data prediction module, the short-term prediction sub-module uses a time series model to predict short-term water resources status, the long-term prediction sub-module performs long-term water resource status prediction, and the trend analysis sub-module performs trend analysis on the prediction results.

In the virtual simulation module, the digital model establishment submodule establishes a digital twin model of the water network, the strategy testing submodule tests different management strategies, and the optimization plan submodule generates optimization plans.

In the anomaly detection module, the leak identification sub-module uses machine learning models to identify leak patterns, the consumption pattern analysis sub-module analyzes abnormal consumption patterns, and the real-time alarm sub-module generates real-time alarm information.

In the notification feedback module, the alarm notification sub-module sends alarm notifications to the maintenance team, the team response sub-module receives the maintenance team's responses and solutions, and the maintenance feedback integration sub-module integrates the maintenance team's feedback information.

Please refer to Figure 3. The pressure sensing sub-module is based on Internet of Things technology and uses a differential pressure detection algorithm to monitor pressure changes in the water in real time, conduct data analysis, and generate real-time water pressure data reports;

Based on the real-time water pressure data report, the flow sensing sub-module uses the turbine flow calculation method to monitor the flow rate and flow rate of the water flow, and combines the water pressure data with the flow estimation to generate a real-time flow data report;

Based on the real-time flow data report, the water quality sensing sub-module uses spectral detection methods to analyze the chemical components in the water, conduct water quality assessment, and generate real-time water quality data reports;

The differential pressure detection algorithm specifically performs differential operations on the water pressure data continuously to obtain the changing trend of the water pressure. The real-time water pressure data report includes pressure values, pressure changing trends and abnormal fluctuations. The turbine flow calculation method specifically calculates the flow rate based on the rotation speed of the turbine. The real-time flow data report includes flow rate, flow value and flow change trend. The spectral detection method is to analyze the spectral characteristics of water samples through spectroscopic instruments. The real-time water quality data report specifically refers to the evaluation report of chemical composition, turbidity, and pH value.

The pressure sensing sub-module is based on Internet of Things technology and uses a differential pressure detection algorithm to monitor pressure changes in the water in real time. First, install the pressure sensor device and connect it with the Internet of Things to achieve remote monitoring and data transmission. Then, the differential pressure detection algorithm is used to continuously perform differential operations on the water pressure data to obtain the changing trend of the water pressure. Finally, a real-time water pressure data report is generated based on the differential results, which includes information such as pressure values, pressure change trends, and abnormal fluctuations.

Based on the real-time water pressure data report, the flow sensing sub-module uses the turbine flow calculation method to monitor the flow rate and flow rate of the water flow, and combines the water pressure data to estimate the flow rate. First, determine the relationship between water flow velocity and flow based on real-time water pressure data reports. Then, install the turbine flow meter device and connect it with the Internet of Things to enable remote monitoring and data transmission. Next, the flow value is calculated based on the rotation speed of the turbine through the turbine flow calculation method. Finally, a real-time flow data report is generated based on the flow rate and flow value, which includes information such as flow rate, flow value, and flow change trend.

The water quality sensing sub-module is based on real-time flow data reporting and uses spectral detection methods to analyze the chemical components in the water and conduct water quality assessment. First, install spectral detection instruments and connect them with the Internet of Things to achieve remote monitoring and data transmission. Then, the spectral characteristics of the water sample are analyzed through spectral detection methods to obtain information on the chemical composition of the water. Next, chemical composition, turbidity, pH, etc. are evaluated based on spectral characteristics and known water quality standards. Finally, a real-time water quality data report is generated, including information such as chemical composition, turbidity, and pH assessment results.

Please refer to Figure 4. The time series data analysis sub-module is based on real-time water affairs data summary and uses a time series prediction algorithm to predict future water pressure changes, formulate adjustment strategies, and generate a water pressure change prediction report;

Based on the water pressure change prediction report, the leak detection sub-module uses anomaly detection algorithms to analyze abnormal drops in water pressure data, determine leak risks, formulate repair plans, and generate leak detection reports;

The pressure adaptive adjustment sub-module is based on the leak detection report and uses fuzzy logic control method to automatically adjust the water pressure according to the leakage risk and consumption pattern, and generate optimized water pressure parameters;

The time series prediction algorithm specifically uses ARIMA or LSTM models to model and predict historical water pressure data. The water pressure change prediction report includes the predicted water pressure value, prediction accuracy and adjustment strategy. The anomaly detection algorithm specifically uses One-Class. The SVM or Isolation Forest method is used to identify abnormal points. The leak detection report includes the abnormal location, leakage degree and emergency repair plan. The fuzzy logic control method specifically makes decisions based on fuzzy sets and fuzzy rules.

In the time series data analysis sub-module, the main formulas of the ARIMA model include the difference, autoregressive part and moving average part of the time series. The modeling process can use the statsmodels library in Python. Here is an example:

from statsmodels.tsa.arima_model import ARIMA,

model = ARIMA(data, order=(p, d, q))，

results = model.fit(),

predictions = results.predict(start=len(data), end=len(data) +forecast_steps - 1, dynamic=False, typ='levels').

LSTM model: LSTM models usually require a deep learning framework such as TensorFlow or PyTorch. Here is an example using TensorFlow:

import tensorflow as tf，

model = tf.keras.Sequential([

tf.keras.layers.LSTM(units=50, return_sequences=True, input_shape=(input_shape, 1)),

tf.keras.layers.LSTM(units=50),

tf.keras.layers.Dense(units=1)

]),

model.compile(optimizer='adam', loss='mean_squared_error')，

model.fit(X_train, y_train, epochs=epochs, batch_size=batch_size)，

predictions = model.predict(X_test).

In the leak detection sub-module, One-Class SVM: One-Class SVM can be implemented using the Scikit-Learn library. The following is the sample code:

from sklearn.svm import OneClassSVM，

model = OneClassSVM(),

model.fit(data)，

predictions = model.predict(data).

Isolation forest: Isolation forest can also be implemented using the Scikit-Learn library. The following is the sample code:

from sklearn.ensemble import IsolationForest,

model = IsolationForest(),

model.fit(data)，

predictions = model.predict(data).

In the pressure adaptive adjustment sub-module, fuzzy logic control can be implemented using fuzzy logic libraries, such as scikit-fuzzy. Here is an example:

import skfuzzy as fuzz，

import numpy as np.

#Create fuzzy sets and membership functions:

pressure = np.arange(0, 101, 1)，

pressure_low = fuzz.trimf(pressure, [0, 25, 50]),

pressure_medium = fuzz.trimf(pressure, [25, 50, 75]),

pressure_high = fuzz.trimf(pressure, [50, 75, 100]).

#Create rule base:

rule1 = fuzz.interp_membership(pressure, pressure_low, input_pressure)，

rule2 = fuzz.interp_membership(pressure, pressure_medium, input_pressure)，

rule3 = fuzz.interp_membership(pressure, pressure_high, input_pressure).

# Fuzzy rules:

aggregated = np.fmax(rule1, np.fmax(rule2, rule3))，

output_pressure = fuzz.defuzz(pressure, aggregated, 'centroid').

Please refer to Figure 5. The multi-dimensional data analysis sub-module is based on real-time water affairs data summary, uses K-means clustering algorithm to conduct dimensional analysis of water quality data, and mines high-frequency problem areas to generate water quality problem clustering results;

The water quality parameter prediction sub-module uses the random forest algorithm to predict the trend of water quality parameters based on the clustering results of water quality problems, and generates water quality parameter prediction results;

The GIS visualization sub-module uses GIS technology to visualize geographical information based on the water quality parameter prediction results, and generates a water quality prediction report;

The K-means clustering algorithm is specifically used to classify water quality data into multi-level quality categories. The random forest algorithm includes integrated learning of decision trees. The water quality parameter prediction results are specifically the water quality change trend in the future time period. The GIS technology is specifically: GIS technology.

The multi-dimensional data analysis sub-module is based on real-time water affairs data summary, uses K-means clustering algorithm to conduct dimensional analysis of water quality data, and mines high-frequency problem areas to generate water quality problem clustering results. First, collect real-time water data, including information such as water quality indicators and geographic location. Then, preprocess the water quality data, including data cleaning and missing value processing. Next, the K-means clustering algorithm is used to classify the water quality data and divide similar water quality indicators into multi-level quality categories. Based on the clustering results, high-frequency problem areas are mined, that is, specific geographical areas where water quality problems occur. Finally, a water quality problem clustering result report is generated, including water quality indicator characteristics of each category and corresponding geographical location information.

The water quality parameter prediction sub-module uses the random forest algorithm to predict the trend of water quality parameters based on the clustering results of water quality problems, and generates water quality parameter prediction results. First, based on the water quality problem clustering result report, water quality indicators related to the problem are selected as prediction targets. Then, a historical data set is prepared, including historical water quality indicators and geographic location information. Next, use the random forest algorithm to train the historical data set and build a prediction model. Enter the geographical location information in the future time period and use the established prediction model to predict the trend of water quality parameters. Finally, a water quality parameter prediction result report is generated, including water quality change trends in various geographical locations in the future time period.

The GIS visualization sub-module uses GIS technology to visualize geographic information based on the water quality parameter prediction results, and generates a water quality prediction report. First, import the water quality parameter prediction results into GIS software. Then, based on the geographical location information, the water quality prediction results of each location are marked on the map. Use the functions provided by GIS software to visually present the water quality prediction results, such as using color mapping to represent different water quality levels or using symbol marks to represent water quality conditions in different geographical locations. Finally, a water quality prediction report is generated, including map display and text description, so that users can intuitively understand the water quality change trend and the distribution of problem areas in the future time period.

Please refer to Figure 6. The remote sensing image analysis sub-module uses the convolutional neural network to extract remote sensing image features based on the remote sensing image data acquired by satellites, and generates a remote sensing image feature report;

The meteorological data processing sub-module uses remote sensing image feature reports and meteorological data, uses data preprocessing methods to clean data, and generates processed meteorological data reports;

The fusion algorithm application sub-module uses deep learning technology to perform fusion analysis of multi-source data based on processed meteorological data reports and real-time water affairs data summary, and generates a comprehensive water affairs analysis report;

Convolutional neural networks are specifically feedforward neural networks and are used for image and sound recognition. Remote sensing image feature reports include feature values and feature vectors extracted from images. Data preprocessing methods include missing value processing, outlier detection and data standardization. Deep learning technology is specifically a multi-layer neural network model, which is used to deal with non-linear relationships.

In the remote sensing image analysis sub-module, the remote sensing image data collected by satellite sensors is first obtained. Subsequently, convolutional neural network (CNN) was used to extract remote sensing image features. This includes building a CNN model that extracts key features such as texture, color, and shape information from images through multiple layers of convolution and pooling operations. Finally, the extracted eigenvalues and eigenvectors are integrated to generate a remote sensing image feature report, which describes in detail the key features extracted from the image.

The meteorological data processing sub-module starts with the data cleaning stage. Combine remote sensing image feature reports and meteorological data to perform data cleaning, including missing value processing, outlier detection and data standardization. This ensures the accuracy and consistency of meteorological data. Subsequently, a processed meteorological data report is generated, which includes statistical information of the cleaned data, missing value processing methods, and outlier detection results.

The fusion algorithm application sub-module is based on processed meteorological data reports and real-time water affairs data summary, and uses deep learning technology, including multi-layer neural network models, to conduct fusion analysis of multi-source data. This process covers the integration of data from multiple sources and modeling nonlinear relationships. Finally, through deep learning technology, a comprehensive water analysis report is generated, including the correlation between data sources, trend analysis, prediction results, etc.

Please refer to Figure 7. The short-term forecast sub-module is based on the comprehensive water analysis report and uses the autoregressive integral moving average model to make short-term forecasts of recent water resources data and generate short-term water resources forecast data;

The long-term prediction sub-module is based on short-term water resources prediction data and uses long short-term memory network to predict medium- and long-term water resources conditions and generate long-term water resources prediction data;

The trend analysis sub-module uses linear regression analysis to conduct water resource development trends and potential risk assessments based on long-term water resources forecast data, and generates water resource forecast reports;

The autoregressive integrated moving average model is used to capture the autoregressive and moving average characteristics of the data. Linear regression analysis specifically uses mathematical methods to model and analyze the linear relationship between variables.

In the short-term forecast sub-module, the required water resources data, such as water level and flow information, are first obtained from the comprehensive water analysis report. Next, the autoregressive integrated moving average model (ARIMA) is used to make short-term forecasts of water resources data in the most recent time period. This includes differencing operations on the data to ensure stationarity, selecting appropriate autoregressive (AR) and moving average (MA) orders, and generating predictions through model training. Finally, the short-term water forecast data generated can be used for practical decision-making and planning.

The long-term prediction sub-module takes short-term water resources prediction data as input. Use the long short-term memory network (LSTM) model to predict medium- and long-term water resources conditions. This includes data preprocessing, LSTM model construction, training and validation. The trained LSTM model is used to generate water resource data prediction results in the future medium and long-term time periods. These results can be used for mid- and long-term water resource planning and management.

The trend analysis sub-module is based on long-term water resources forecast data. First, the development trends and potential risks of water resources are modeled through linear regression analysis. This includes selecting an appropriate linear regression model, fitting the data, assessing the significance of the regression coefficients, and interpreting the model. With the help of the established linear regression model, the future development trend of water resources is predicted. Finally, a water resources forecast report is generated, which includes forecast results of trends, assessment of potential risks, and other relevant information for reference by decision-makers and stakeholders.

Please refer to Figure 8. The digital model establishment sub-module uses three-dimensional modeling technology to establish a digital water network model based on the water resources prediction report and generate a digital water network model;

The strategy testing sub-module is based on the digital water network model and uses Monte Carlo simulation method to test the effect of water resources management strategies and generate strategy test results;

Based on the strategy test results, the optimization plan sub-module uses the decision tree analysis method to propose the optimal management strategy and optimization plan, and generates a simulation test report;

Three-dimensional modeling technology specifically refers to the use of computer-aided design software to create geometric representations of spatial objects. Monte Carlo simulation methods simulate and analyze future results by randomly extracting parameters from probability distributions. Decision tree analysis methods specifically refer to decision support tools, using Tree diagrams and prediction results to analyze probabilistic event outcomes, resource costs and benefits.

First, in the digital model establishment sub-module, key data in the water resources prediction report are collected, such as water level, flow, water quality and other information. Then, computer-aided design software is used to construct a digital water network model with three-dimensional modeling technology, including pipelines, pumping stations, pools and other elements, and is associated with water resources data to generate a digital water network model for subsequent strategy testing. and provide the basis for optimization.

Next, enter the strategy testing submodule. Use Monte Carlo simulation methods to evaluate the effects of different water resource management strategies. First, parameters are randomly selected to simulate different future scenarios, involving the randomization of parameters such as different meteorological conditions, water resource needs, etc. Then, through Monte Carlo simulation, the effects of water resources management strategies under different scenarios are simulated, including changes in water level, water quality, water supply capacity, etc. Finally, strategy test results are generated, including water resource status and effect assessments under different scenarios, to evaluate the feasibility of various strategies.

Based on the strategy test results, the optimization plan sub-module uses the decision tree analysis method to propose the optimal water resources management strategy and optimization plan. First, organize the strategy test results and related data, including information on the effectiveness and cost of different strategies. Decision trees are then constructed using decision tree analysis methods to help decision makers understand the potential outcomes and trade-offs of different strategies. In this process, optimal water resources management strategies and optimization plans are proposed, taking into account resource costs and benefits. Finally, a simulation test report is generated, which includes the proposed optimal strategy, a detailed description of the optimization plan, and an assessment of potential risks.

Please refer to Figure 9. Based on the comprehensive water analysis report, the leakage identification sub-module uses a convolutional neural network to analyze water pipeline image data, identify leakage characteristics, and generate a leakage identification report;

The consumption pattern analysis sub-module uses the K-means clustering algorithm to analyze consumer water usage data based on leakage identification reports, detect abnormal consumption patterns, and generate abnormal consumption pattern reports;

The real-time alarm sub-module is based on abnormal consumption pattern reporting and uses threshold analysis method for real-time data monitoring. When the data exceeds the preset threshold, an alarm is triggered and an abnormal event report is generated;

The convolutional neural network specifically uses deep learning technology to extract features from pipeline system images and identify leak locations and sizes. The K-means clustering algorithm includes clustering user consumption data and dividing user groups based on multi-dimensional data including water consumption and time. , to identify abnormal consumption patterns. The threshold analysis method specifically refers to setting the safe range of key indicators of water flow and pressure. When the real-time data exceeds the safe range, an alarm is automatically triggered.

In the leak identification sub-module, firstly, a comprehensive water analysis report is obtained from the water system and the water pipeline image data is extracted. These image data are then preprocessed, including steps such as denoising and enhancement, to improve the recognizability of leakage features. Then, the deep learning technology of convolutional neural network (CNN) is used to analyze the preprocessed image data. CNN is able to extract features in images to help accurately identify the location and size of leaks. Finally, based on the CNN analysis results, a leak identification report is generated, describing the characteristics of the leak in detail, including the location, size, and severity of the leak.

In the consumption pattern analysis sub-module, leakage identification reports are used as input data, and consumers' water use data are collected at the same time, including multi-dimensional information such as water consumption and time. These data are preprocessed, cleaned and standardized in preparation for subsequent K-means clustering algorithm. K-means clustering algorithm is applied to consumption data to classify users into different groups and detect abnormal consumption patterns based on water use characteristics and temporal patterns. Finally, an abnormal consumption pattern report is generated, which includes detection results of abnormal water usage patterns and detailed information about which users have abnormal water usage behaviors.

In the real-time alarm sub-module, abnormal consumption pattern reports are used as input data, and data on key indicators such as real-time water flow and pressure are obtained from sensors or monitoring equipment. This real-time data is subjected to threshold analysis and compared with pre-set safety limits. When real-time data exceeds the safe range, the threshold analysis method triggers an alarm. The triggered alarm information includes the time, location and related abnormal indicators of the abnormal event, and is integrated into the abnormal event report.

Please refer to Figure 10. The alarm notification sub-module uses automatic message push technology to notify the maintenance team of alarms based on abnormal event reports and simulation test reports, and generates alarm notification records;

The team response sub-module uses instant messaging software to coordinate within the team based on alarm notification records, formulate emergency response plans, and generate team response records;

The maintenance feedback integration sub-module is based on team response records, uses data fusion technology to integrate maintenance feedback, including problem solving progress and effects, and generates maintenance feedback records.

In the alarm notification sub-module, first, using abnormal event reports and simulation test reports, the system automatically identifies abnormal events and generates alarm notifications. These notifications are sent to maintenance team members via automated push messaging technology to notify them of potential issues. At the same time, the system records relevant information about alarm notifications, including time, event type, location, etc., and generates alarm notification records for subsequent tracking and review.

In the team response sub-module, after team members receive the alarm notification, they use instant messaging software for internal coordination and communication. Together they develop an emergency response plan, including identifying the measures that should be taken, the division of labor and the timetable. This process is recorded and a team response record is generated, which includes information such as the developed response plan, key decisions, and responsibility assignments for subsequent tracking and review.

In the maintenance feedback integration sub-module, based on team response records, data fusion technology is applied to integrate maintenance feedback information into the system. This includes the progress, effectiveness, and any other relevant information of problem resolution. This data is integrated into maintenance feedback records to provide a comprehensive view of problem resolution and the effectiveness of the maintenance process. This can be used to continuously improve maintenance processes and ensure system stability and reliability.

The above are only preferred embodiments of the present invention and do not limit the present invention in other forms. Any skilled person familiar with the art may use the technical content disclosed above to make changes or modifications to equivalent embodiments with equivalent changes. In other fields, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the technical content of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims (10)

1. Intelligent water affair partition metering system based on internet big data analysis, its characterized in that: the intelligent water service partition metering system based on internet big data analysis comprises a data collection module, a water pressure adjusting module, a water quality monitoring module, a multi-source data fusion module, a data prediction module, a virtual simulation module, an abnormality detection module and a notification feedback module;

The data collection module is based on the internet of things technology, adopts sensor equipment to collect real-time water service data, and comprises water pressure, flow and water quality information to generate real-time water service data summary;

the water pressure regulating module is used for carrying out water pressure leakage and consumption mode analysis by adopting a neural network model based on real-time water service data summarization, carrying out water pressure self-adaptive regulation and generating optimized water pressure parameters;

the water quality monitoring module predicts water quality change by adopting a random forest algorithm based on real-time water affair data summarization, and simultaneously performs data visualization by utilizing a GIS (geographic information system) to generate a water quality prediction report;

the multi-source data fusion module combines real-time water affair data summarization with satellite and meteorological data, adopts a deep learning technology to carry out data fusion, carries out water affair analysis and generates a comprehensive water affair analysis report;

the data prediction module predicts the water resource condition based on the comprehensive water affair analysis report by using a time sequence model, performs trend analysis and generates a water resource prediction report;

the virtual simulation module is used for carrying out water network simulation by utilizing a digital twin technology based on the water resource prediction report, carrying out management strategy test and generating a simulation test report;

The anomaly detection module is used for carrying out anomaly consumption and leakage pattern recognition by adopting a machine learning model based on the comprehensive water affair analysis report and generating an anomaly event report;

the notification feedback module is used for carrying out maintenance team notification and problem feedback collection based on the abnormal event report and the simulation test report, and generating a maintenance feedback record;

the real-time water service data summarizing specifically is multi-node water pressure, flow and water quality data stored in a time sequence form, the water pressure data comprise temperature, pH value and turbidity, the optimized water pressure parameters specifically are water pump operation parameters adjusted based on consumption modes and leakage conditions, the water quality prediction report specifically predicts the water quality change trend of multiple nodes in a future time period, the comprehensive water service analysis report comprises a pump station operation condition, a water pipe network structure, water quality conditions and a user consumption mode, the water resource prediction report specifically is the predicted change and demand trend of water resources in the future time period, the simulation test report specifically is a test result and an optimization scheme, and the abnormal event report specifically is real-time pipe network leakage data and abnormal consumption modes.

2. The intelligent water service partition metering system based on internet big data analysis according to claim 1, wherein: the data collection module comprises a pressure sensing sub-module, a flow sensing sub-module and a water quality sensing sub-module;

The water pressure regulating module comprises a time sequence data analyzing sub-module, a leakage detecting sub-module and a pressure self-adaptive regulating sub-module;

the water quality monitoring module comprises a multidimensional data analysis sub-module, a water quality parameter prediction sub-module and a GIS visualization sub-module;

the multi-source data fusion module comprises a remote sensing image analysis sub-module, a meteorological data processing sub-module and a fusion algorithm application sub-module;

the data prediction module comprises a short-term prediction sub-module, a long-term prediction sub-module and a trend analysis sub-module;

the virtual simulation module comprises a digital model building sub-module, a strategy testing sub-module and an optimizing scheme sub-module;

the abnormality detection module comprises a leakage identification sub-module, a consumption mode analysis sub-module and a real-time alarm sub-module;

the notification feedback module comprises an alarm notification sub-module, a team response sub-module and a maintenance feedback integration sub-module.

3. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the pressure sensing submodule monitors pressure change in water in real time by adopting a differential pressure detection algorithm based on the internet of things technology, performs data analysis, and generates a real-time water pressure data report;

The flow sensing submodule monitors the flow speed and the flow of water flow by adopting a turbine flow calculation method based on the real-time water pressure data report, carries out flow estimation by combining the water pressure data, and generates a real-time flow data report;

the water quality sensing submodule analyzes chemical components in water by adopting a spectrum detection method based on the real-time flow data report, and carries out water quality assessment to generate a real-time water quality data report;

the differential pressure detection algorithm is characterized in that differential operation is continuously carried out on water pressure data to obtain a variation trend of water pressure, the real-time water pressure data report comprises a pressure value, a pressure variation trend and abnormal fluctuation, the turbine flowmeter algorithm is used for calculating a flow value according to the rotating speed of a turbine, the real-time flow data report comprises a flow speed, a flow value and a flow variation trend, the spectrum detection method is used for analyzing the spectrum characteristics of a water sample through a spectrum instrument, and the real-time water quality data report is used for evaluating and reporting chemical components, turbidity and pH value.

4. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the time sequence data analysis sub-module predicts future water pressure changes based on real-time water service data summarization by adopting a time sequence prediction algorithm, and carries out regulation strategy formulation to generate a water pressure change prediction report;

The leakage detection submodule analyzes abnormal decline in the water pressure data by adopting an abnormal detection algorithm based on the water pressure change prediction report, determines leakage risk, and formulates a repair scheme to generate a leakage detection report;

the pressure self-adaptive adjusting submodule automatically adjusts the water pressure according to leakage risks and consumption modes by adopting a fuzzy logic control method based on the leakage detection report to generate optimized water pressure parameters;

the time sequence prediction algorithm is specifically to model and predict historical water pressure data by adopting an ARIMA or LSTM model, the water pressure change prediction report comprises a predicted water pressure value, prediction accuracy and an adjustment strategy, the anomaly detection algorithm is specifically to identify anomaly points by adopting an One-Class SVM or Isolation Forest method, the leakage detection report comprises an anomaly position, leakage degree and an emergency repair scheme, and the fuzzy logic control method is specifically to make decisions based on a fuzzy set and a fuzzy rule.

5. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the multidimensional data analysis submodule performs dimensional analysis on water quality data by adopting a K-means clustering algorithm based on real-time water service data summarization, and digs a problem high-frequency region to generate a water quality problem clustering result;

The water quality parameter prediction submodule predicts the trend of the water quality parameter by adopting a random forest algorithm based on the water quality problem clustering result to generate a water quality parameter prediction result;

the GIS visualization submodule performs geographic information visualization presentation by utilizing a GIS technology based on a water quality parameter prediction result to generate a water quality prediction report;

the K-means clustering algorithm is used for classifying water quality data into multilevel quality categories, the random forest algorithm comprises integrated learning of decision trees, the water quality parameter prediction result is a water quality change trend in a future time period, and the GIS technology is a geographic information system technology.

6. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the remote sensing image analysis submodule is used for extracting the characteristics of the remote sensing image by adopting a convolutional neural network based on the remote sensing image data acquired by the satellite, and generating a remote sensing image characteristic report;

the meteorological data processing submodule carries out data cleaning by adopting a data preprocessing method based on the remote sensing image characteristic report and meteorological data to generate a processed meteorological data report;

The fusion algorithm application submodule carries out fusion analysis of multi-source data by adopting a deep learning technology based on the processed meteorological data report and real-time water service data summary to generate a comprehensive water service analysis report;

the convolutional neural network is specifically a feedforward neural network and is used for image and sound identification, the remote sensing image characteristic report comprises characteristic values and characteristic vectors extracted from images, the data preprocessing method comprises missing value processing, abnormal value detection and data standardization, and the deep learning technology is specifically a multi-layer neural network model and is used for processing nonlinear relations.

7. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the short-term prediction submodule carries out short-term prediction on the recent water resource data by adopting an autoregressive integral moving average model based on the comprehensive water analysis report to generate short-term water resource prediction data;

the long-term prediction sub-module predicts the medium-term and long-term water resource conditions by adopting a long-term and short-term memory network based on the short-term water resource prediction data to generate long-term water resource prediction data;

the trend analysis sub-module is used for carrying out development trend and potential risk assessment of the water resource by adopting linear regression analysis based on the long-term water resource prediction data to generate a water resource prediction report;

The autoregressive integral moving average model is used for capturing autoregressive and moving average characteristics of data, and the linear regression analysis is specifically used for modeling and analyzing linear relations among variables by using a mathematical method.

8. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the digital model building sub-module is used for building a digital water network model by adopting a three-dimensional modeling technology based on the water resource prediction report to generate the digital water network model;

the strategy testing submodule is used for performing effect testing of the water resource management strategy by adopting a Monte Carlo simulation method based on the digital water network model to generate a strategy testing result;

the optimizing scheme sub-module adopts a decision tree analysis method to propose an optimal management strategy and an optimizing scheme based on a strategy test result, and generates a simulation test report;

the three-dimensional modeling technology specifically refers to the use of computer aided design software to create geometric representation of a space object, the Monte Carlo simulation method performs simulation analysis on future results by randomly extracting parameters from probability distribution, the decision tree analysis method specifically refers to a decision support tool, and probability event results, resource cost and benefit are analyzed by using a tree diagram and a prediction result.

9. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the leakage identification submodule analyzes water pipeline image data based on the comprehensive water analysis report by adopting a convolutional neural network, identifies leakage characteristics and generates a leakage identification report;

the consumption mode analysis submodule analyzes the water use data of the consumers by adopting a K-means clustering algorithm based on the leakage identification report, detects an abnormal consumption mode and generates an abnormal consumption mode report;

the real-time alarm submodule monitors real-time data by adopting a threshold analysis method based on the abnormal consumption mode report, triggers an alarm when the data exceeds a preset threshold, and generates an abnormal event report;

the convolutional neural network specifically utilizes a deep learning technology to perform feature extraction on a pipeline system image and identify the leakage position and the leakage size, the K-means clustering algorithm comprises a user consumption data cluster, a user group is divided according to multidimensional data comprising water consumption and time, an abnormal consumption mode is identified, the threshold analysis method specifically refers to setting the safety range of key indexes of water flow and pressure, and when real-time data exceeds the safety range, an alarm is automatically triggered.

10. The intelligent water service partition metering system based on internet big data analysis according to claim 2, wherein: the alarm notification sub-module carries out alarm notification on a maintenance team by adopting an automatic message pushing technology based on the abnormal event report and the simulated test report, and generates an alarm notification record;

the team response submodule is used for carrying out team internal coordination by utilizing instant messaging software based on the alarm notification record, making an emergency response scheme and generating a team response record;

the maintenance feedback integration submodule integrates maintenance feedback, including the problem solving progress and effect, based on the team response record by applying a data fusion technology, and generates a maintenance feedback record.