Keywords

1 Introduction

China is a big country of coal production and consumption in the world. Coal is mainly used in the four industries of power, metallurgy, building materials and chemical industry, among which the proportion of the power industry is the largest. The coal and electricity energy supply chain consists of coal production, coal transportation, power generation, power transmission, and power utilization [1].

Since China’s electricity is mainly derived from coal power generation, it has important influence on the supply and demand of the whole coal market. At present, there are many researches on the separate part of the thermal coal supply chain at home and abroad, but there is less research on the overall price risk. As a link between electricity and coal, the power coal can truly reflect the relations between the two industries. So it is important to study the thermal coal supply chain and deeply analyze factors that affect the thermal coal demand for power generation companies, including power generation plans, inventory, coal prices, transportation costs, supply cycle, etc. It can help them predict thermal coal demand, rational arrange fuel purchase plans and optimize business strategy. In addition, the selection of suppliers also plays an important role in supply chain management. It directly affects the competitiveness of power companies in the entire supply chain. Therefore, it is necessary to portray corporate portraits for coal suppliers by big data technology. Through corporate portraits, power companies can find suitable coal suppliers in terms of service, delivery time, coal prices, and transportation costs.

2 System Design

With the reform of organization in electric power industry, the operational mechanism and the management methods of power supply enterprise have great changes. Electric power enterprises are going into market economy. If the power generation enterprises do not obtain and analyze the data rooted in production and management, realize cost reduction and efficiency enhancement, grasp the needs of customers, and understand marketing strategies of competitors, they will lose their initiative and core competitiveness in the reform. Therefore, we designed SCTC based on big data and visualization technology. SCTC can assist power enterprises in predicting their own coal demand within a limited time, and can quickly select suitable partners from neighboring coal companies when purchasing coal. So the system mainly implements the following three functions:

  • Coal demand forecast

  • Regional situation analysis

  • Enterprise portraits of coal suppliers.

3 Technical Route

3.1 System Architecture

In the paper, we designed and implemented a visual system for the supply chain of thermal coal, which is analyzed and displayed from multi-dimensional. It facilitates decision makers to compare and analyze coal suppliers in a short period of time. The system uses Echarts data visualization technology to achieve a flexible, intuitive and interactive chart. At the same time, it also provides functions such as data views, area selection, multi-map linkage, and sub-area maps. The user can mine and integrate data in the system.

The system is based on the user request response framework. It mainly consists of the following three parts: database, server and Web client. The client sends a HTTP request to the server first, and the front controller distributes it to the corresponding page controller according to the request information (such as URL). After receiving the request, the page controller delegates the request object to the business object for processing, interacts with the database through MyBatis, and implements data processing. After the end of the process, the data and view are returned to the front end controller. Then the front controller reclaims control, hands over data and views to Vue.js, and Vue.js binds the data to Echarts. Finally the front controller presents the data and the page to the user [2]. Its framework is shown in Fig. 1.

Fig. 1.
figure 1

The framework of SCTC

Full size image

3.2 Data Preparation and Processing

The data used in the paper is obtained through network crawlers on many networks. These data are divided into three main categories: coal-related data, coal enterprise related data and power generation enterprise data. Figure 2 shows the classification of a coal network data after processing.

Fig. 2.
figure 2

The classification of a coal network data

Full size image

We acquired coal-related data such as coal price, coal quality, production volume, inventory, transportation cost, and supply cycle, and related coal enterprise data such as company size, corporate background, cooperation, corporate credit, and relationship network, and the location information of the coal supply enterprise through the web crawler. In addition, we also obtained data related to power companies of the Group’s subsidiaries, such as geographical location, power generation plan, inventory, and so on. After preparing the data, the system will store them in the database.

3.3 Coal Demand Forecast

SCTC provides the forecasting function of coal demand, using Auto-Regressive and Moving Average model (ARMA) model and Multivariate Co-integration Analysis model. It can forecast the consumption of electric coal for power companies [3].

3.3.1 Model Introduction

ARMA Model

ARMA model is a combination of Auto-Regressive model (AR) and Moving-Average model (MA). Therefore, it is also called autoregressive moving average model [4]. Its equation is as follows:

$$ x_{t} = c + \emptyset_{1} x_{t - 1} + \cdots + \emptyset_{p} x_{t - p} + u_{t} + \theta_{1} u_{t - 1} + \cdots + \theta_{q} u_{t - q} ,\;t = 1,2, \cdots {\text{T}} $$
(1)

In the equation, C is a constant. \( \emptyset_{\text{i}} \) and \( \emptyset_{\text{j}} \) represent the coefficients of the Auto-Regressive model and the Moving-Average model, respectively. p and q represent the order of the two models.

Multivariate Co-integration Analysis Method

Co-integration theory was proposed by Engle and Granger in 1987. It is the basis for the study of dynamic relationships.

Assume that the sequence of independent variables and the sequence of dependent variables are \( \left\{ {{\text{X}}_{1} } \right\}, \cdots \left\{ {{\text{X}}_{\text{k}} } \right\} \) and \( \left\{ {y_{t} } \right\} \) respectively. Construct a regression model as follows [5].

$$ y_{t} = \beta_{0} + \sum\nolimits_{i = 1}^{k} {\beta_{i} \,x_{it} + \varepsilon_{t} } $$
(2)

If the test result shows that \( \left\{ {\upvarepsilon_{\text{t}} } \right\} \) is stationary, it means that there is a cointegration relationship between the sequence \( \left\{ {{\text{y}}_{\text{t}} } \right\} \) and \( \left\{ {{\text{X}}_{1} } \right\}, \cdots \left\{ {{\text{X}}_{\text{k}} } \right\} \).

3.3.2 Function Introduction

In the system, when the function of coal demand forecast is used, users can choose the geographical location including the whole country and the provinces and cities. When a user searches for a power plant within the scope of its own authority, the system can display historical data on a monthly basis and predict energy production and consumption of the plant in the coming months.

In SCTC, a two-column display is used for the planned and actual production capacity, and the predicted consumption and actual consumption are displayed with two-color columns of another color. When the mouse is hovered over the two-column column, the proportion prompt will pop up. The function enables business staff to more intuitively understand the plan and the dynamics of actual production, and the ratio of the deviation between the predict consumption and the actual consumption. At the same time, it also helps the system developers to collect data, optimize the algorithm, and improve the accuracy of the system prediction.

3.4 Regional Situation Analysis

The regional situation analysis interface uses the layout of the middle main view and the surrounding auxiliary view. The main view and the auxiliary view can be linked together.

The main view is a GIS map that allows users to select power generation companies within the country or provinces or cities. After selecting a power generation company on the map, the coal suppliers around the power generation company will be immediately displayed on the map, and the distance between the suppliers and the power generation companies and the corresponding transportation costs will be marked.

The auxiliary view includes the overview of the electric power enterprise, the coal price, the supply cycle, the business operation, the coal quality contrast, the enterprise credit and the coal price trend of the top 5 coal enterprises. The overview of the power generation company mainly shows basic information such as the power generation plan, coal inventory, and optimal coal quality of the company. The coal quality section includes information such as calorific value, sulfur content, moisture content, ash content, and comprehensive evaluation of the coal enterprises. In the form of the linkage between the main view and the auxiliary view, the user can select a suitable coal supplier in a short time.

3.5 Enterprise Portraits of Coal Suppliers

In the regional situation analysis interface, the user can grasp the situation of all the coal suppliers in a specific area, and can initially filter out several suitable suppliers in less time. In order to grasp the full range information of coal suppliers, SCTC provides enterprise portraits. When a coal supplier is selected in the regional situation interface, it will automatically jump to the enterprise portraits interface of the coal supplier company.

The portraits of coal companies are mainly described in terms of basic information, evaluation and relations, and business cooperation. The basic information includes the scale of the company, business background, production capacity, coal quality, and coal price. The evaluation and relations include corporate credit evaluation, supplier evaluation, and relationship network. The business cooperation includes contract conditions, cohesiveness, coal purchases, and coal purchases. Three-color labels represent the three states above, equal to, and below the average of all companies for quantifiable indicators. The relationship network is more complex, so it is shown in thumbnails. An individual display window for the relational network pops up when the thumbnail is clicked.

The intelligent analysis part uses big data technology to clean and mine target enterprise data. The comprehensive assessment of coal companies is conducted in terms of transaction volume in the past year, changing trend of historical trading average prices, credit evaluation and so on.

4 The Application

Through the analysis and mining of coal consumption, coal companies and power generation enterprises and other related data in February 2017, the paper forecasts the coal demand of a power generation enterprise, analyzes coal suppliers of the power enterprise, and provides the portrait of a coal supplier company. The results are shown in Figs. 3, 4, and 5.

Fig. 3.
figure 3

A coal demand of a power company

Full size image
Fig. 4.
figure 4

Regional situation analysis

Full size image
Fig. 5.
figure 5

The enterprise portrait of a coal supplier

Full size image

The top five coal companies that are most suitable for this power generation company to purchase coal in Inner Mongolia, and as shown in Fig. 4.

5 Conclusion

The paper introduces the design and application of the visual system for the supply chain of thermal coal (SCTC), and implements the functions of coal demand prediction, regional situation analysis and the corporate portrait of the coal supplier. It has been verified by the actual data. The system is simple and friendly. Using the system, users do not have to compare information about coal prices, coal quality, supply time and so on by artificial. SCTC reduces the mistakes that can be made when looking for information subjectively, provides help and support for management and decision-making on purchase of thermal coal.