The system consists of two parts: software and hardware platforms14. The software platform is a functional module based on a B/S architecture that conducts data acquisition, data analysis and processing, data visualization, data trend analysis, air pollutant interpolation analysis, and data early warning. The hardware platform is a wireless sensor network, which includes a motherboard and data acquisition board.
The software operating environment is a Web portal server with 4 CPUs, 8 GBs of memory, 160 GBs of disk, and the 2012 Windows Server operating system. The software environment has a running environment that uses JAVA 8, Apache Tomcat 9.0.8, MySQL, and other software preinstalled on the server.
Hardware platform design
Since the platform is a sensory platform, the hardware system adopts a modular design, which was mainly applied to the design and development of the motherboard and data acquisition board. With an ARM processor as the core, the IPv6 protocol is embedded in each sensor node to operationalize the data acquisition and upload functions.
Motherboard
The motherboard is the core component of a wireless sensor, and it is composed of a charging management unit, power conversion unit, clock management unit, network interface, J-Link component, data storage unit, network interface, and signal transceiver. With the STM32F107 chip as the core, the IPv6 protocol is embedded to control the acquisition and storage of data in field environments and is responsible for the communication between wireless sensors and servers.
Data acquisition board
Air quality is monitored by a series of air pollutant concentration indexes, so equipment must be used to connect different sensors and equipment15. For different sensor types, the acquisition board requires various connection modes to interact with the motherboard. Therefore, the data acquisition board is separated from the motherboard to improve the universality and scalability of the whole system. The separately designed digital acquisition board connects to each type of sensor. For example, a digital sensor is the communications interface between the sensor and the processor provided by the data acquisition board; an analog sensor is a functional component provided by the data acquisition board to connect the final amplified or converted data with the processor when the sensor signal needs to be amplified and converted, and the output pulse sensor is used to shape the pulse signal output by the sensor. The I/O ports on the motherboard include 12 A/D converters and 12 D/A converters, as well as synchronous and asynchronous transceivers, a controller area network, timers, counters, and internal integrated circuits16. The entire hardware framework is shown in the following Fig. 1.
Figure 1
The entire hardware framework.
On the platform, the data acquisition board is shown in Fig. 1. Through this board, the sensor can communicate with the mainboard and provide data interaction. For the above-mentioned different types of sensors, the interaction between the data acquisition board and the mainboard is different. For example, for digital sensors, the data acquisition board only provides an interface for these sensors to communicate with the motherboard. For analog sensors, the data acquisition board amplifies and converts the electrical signals output by the sensors, and then passes them to the mainboard for processing through a dedicated interface.
Software platform design
The overall framework of the monitoring and early warning platform for air pollutants with high space–time accuracy based on IPv6 is shown in Fig. 1. The platform is divided into the data acquisition module, the data analysis, and processing module, and the data visualization and early warning module.
The data acquisition module monitors the status of atmospheric pollutants through wireless sensors, builds a system database, and sends the data on the atmospheric pollutant concentration status to the background server in real-time. The data analysis and processing module are responsible for receiving the message data sent by the wireless sensor and unpacking and analyzing the data according to the specified analysis rules. The data visualization and early warning module is responsible for the front-end display of the data, determines abnormal monitoring statuses, flags unexpected monitoring data from the monitoring points based on the received data, and feeds the information back to the management personnel in a timely fashion.
Data acquisition module
The design and implementation of the data acquisition module require a 4G communication network to implement data uploading, and a wireless sensor network collector terminal implements data forwarding through an embedded system. At present, the socket port monitoring data on the server are implemented using the TCP protocol17. After the terminal data collector collects data, it requests that the server interface establish a connection with the server to facilitate communication. After the server receives the data transmitted by the collector, it unpacks and analyzes the collected data according to the format of the protocol message. For a specific acquisition factor, it is necessary to analyze the data according to their corresponding analysis parameters. After the analysis is completed, the data are stored in the database to wait for the data call. The specific operation flow is as follows:
-
1
First, the data are accepted and stored in the local buffer through the socket interface, and the completely agreed message data are accepted
-
2
Then, the sensor mark position is analyzed corresponding to the data, and the data are converted behind the data position, according to whether the corresponding sensor has data (0 means no data, 1 means data)
-
3
According to the agreed resolution factor of each acquisition parameter, different parameters are multiplied by different resolution factors, and then, the data-bind with the equipment after the corresponding equipment ID is resolved.
-
4
To ensure the readiness of the data, the calibration equipment is installed at the near-ground position, and the accuracy of the data collected by the sensors is updated by comparing the collected and analyzed data.
Data analysis and processing module
The data analysis and processing module analyze the collected data, and the current TCP connection node is message data obtained through the socket interface. For different acquisition factors, data are arranged according to whether the corresponding data collector has data, and then, the data format is filled after the flag bit. The data bits have already agreed to the arrangement rules in the data analysis part, and each bit represents a different data sensor. To save the stored message length, the collected data are filled after the flag bit; otherwise, they are not filled.
As shown in above Table 3, the collected data, the data flag bits, and the corresponding data bits are analyzed to obtain the corresponding real data.
Table 3 The specific meaning of each position of the received data.
To facilitate the transmission of data and to avoid the complexity of data formats caused by decimals, the scaling parameters of each parameter are agreed upon through the parsing program during data acquisition to ensure that the data transmission results are integers. After unpacking and parsing the basic data, the data are post-processed according to the agreed-upon parameters. The corresponding acquisition equipment binds to the data according to the transmitted equipment number.
Data visualization
After collecting the high spatiotemporal precision data collected by the wireless sensor network in the server, our main task is to visualize the data to show changes in an atmospheric environment in a useful way for decision-makers.
The overall architecture of data visualization is shown in the figure below, which is mainly composed of a front-end call, background data processing, and a data storage module. For the background data call process, we use the spring MVC framework in Java, which ensures the implementation of the front-end real-time data display function18. Figure 2 shows the statistical interface of the latest data for each site:
Figure 2
The statistical interface of the latest monitoring data of each site of the monitoring and early warning platform.
Data trend analysis and change
The data trend analysis changes of the stations in the experimental area include hourly trend analyses, daily average trend analyses, and data trend analyses for each station. The latest collected data are displayed in node charts with high spatial and temporal distributions installed in the area. For the collected data, the air pollution index is calculated corresponding to each collection time based on the calculation method stipulated by its state. The ranking of data collections at multiple points in the region is calculated according to the air pollution index, and the pollution situation of different observation points is analyzed. Based on the installation nodes, the API function of the network map is used to distribute the nodes on the map, and then, the distribution of the collection sites is obtained from the entire area. The following Fig. 3 shows a heat diagram of the change trends of different observation parameters for different dates:
Figure 3
Because of the AQI, PM2.5, PM10, SO2, NO2, CO, different observation parameters such as O3 change trend in different periods to heat (From: Bai du map).
Interpolation analysis of air pollutants
The toolbox based on ArcGIS interpolates the observation point data with high spatial and temporal distributions and then obtains the air pollution status of the entire region through limited data collection points. Using the collected data, a different analysis of the regional fine particles, nitrogen and sulfur oxides, the environmental pollution index can be obtained. By using time series data, continuous monitoring of regional environmental pollution can be performed, and by including the monitoring of unexpected data, timely reporting and decision-making for problematic pollution are enabled.
After collecting the different data, we can calculate the data of different places. Via comparison with data from known observation points, we can choose a more suitable interpolation algorithm. By calling the interpolation algorithm interface of the ArcGIS server, we can find an interpolation algorithm with the lowest error for mapping.
Data early warning module
The data early warning module is divided into abnormal data collections and abnormal data thresholds. As shown in Fig. 4. The figure shows that the device named SH had a machine failure at 14:46:07 on 2018.05.21.
Figure 4
The platform early warning module includes data acquisition anomalies and data threshold anomalies.
A data collection anomaly triggers an abnormal alarm if the remote collector receives an abnormal collection (that is, no data are collected); then, the system will push the information to the person in charge, and the appropriate personnel will carry out the specific equipment maintenance.
The principle of the data threshold alarm is to set a specific monitoring threshold for specific pollutants. If the threshold is exceeded, then early warning information is triggered, allowing problematic pollution incidents to be found and treated in time. The data threshold anomaly is divided into two parts. One part makes a judgment according to a set threshold. System administrators can set different early warning thresholds for different pollutants, which will trigger data alarms when the collected data exceed the current threshold (the corresponding situation may be a sudden increase in environmental pollution). The other part is an early warning for different pollutant variation ranges, which are triggered based on early warning information when the collected data increase beyond a certain range. On the whole, the data early warning system mainly relies on the high spatial and temporal distribution data of the wireless sensor networks. This system implements real-time monitoring of pollutant concentration changes for specific observation points in an entire area and processes decisions for sudden high pollution events in a timely fashion.
[ad_2]
Originally Appeared Here