JP7518591B2 - Process management support device, support method, support program, and system - Google Patents

Description

The present invention relates to a support device, support method, support program, and system for process management in a cement manufacturing plant.

In recent years, efforts have been underway in the process industry to further improve plant operations using AI. The cement industry is no exception, and efforts are being made to expand the scope of its application, including quality prediction systems.

In process industries such as the cement industry, process management is carried out by operators judging the situation from multiple process data and selecting an operation. The operator's judgment includes the selection of the operation target, the amount of operation, and the timing, and there are various patterns, so the operating state depends on the operator's individual skill. Furthermore, such skills are difficult to convey as objective information, and their transfer is extremely difficult.

To improve this situation, technology has been proposed for transferring work skills in plant operations. For example, the operation support device described in Patent Document 1 acquires various types of information, extracts the relationships between the various types of information, and provides work procedures and the like that support plant operation based on the extracted relationships.

On the other hand, a plant operation support system that displays three-dimensional information to support operation is known. For example, the operation support system described in Patent Document 2 displays three-dimensional information on three axes for an energy supply facility that has multiple energy forms, the time, the target operation pattern corresponding to the peak of the target operation pattern probability density distribution, and the target operation pattern probability density distribution.

JP 2019-3545 A JP 2003-143757 A

However, at plant work sites, support is required according to the ever-changing situation, and this cannot be achieved by simply providing standard work procedures as in the system described in Patent Document 1. The system described in Patent Document 2 displays a target operating pattern according to the time, but does not display the progress of the situation related to product quality.

The present invention has been made in consideration of these circumstances, and aims to provide a process management support device, support method, support program, and system that can clearly communicate to operators not only the current status of the manufacturing process, but also future predicted status regarding product quality.

(1) In order to achieve the above object, the process management support device of the present invention is a process management support device in a product manufacturing plant, and is characterized by comprising: an information acquisition unit that acquires process data representing the environment or settings in the target plant in real time; a predicted value identification unit that identifies a predicted value for a predetermined parameter in the acquired process data using a layer structure that has been deep-learned using an accumulated data set of the process data; and a display control unit that indicates the direction in which the predetermined parameter should be changed by displaying the predetermined parameter and the predicted value in the acquired process data superimposed on a display of a predetermined standard related to product quality.

This allows the operator to easily understand not only the current status of the manufacturing process, but also future predicted status regarding product quality. As a result, the operator can utilize the provided information in their operations to improve the manufacturing process.

(2) The process management support device of the present invention is also characterized in that it displays the predetermined criteria, the predetermined parameters of the acquired process data, and the predicted values on an xy plane with xy axes formed by the two types of the predetermined parameters. This makes it possible to display the direction in which the process data is heading in an intuitive and easy-to-understand manner.

(3) Furthermore, the process management support device of the present invention is characterized in that the predetermined criterion is represented by a closed figure on the xy plane, and the direction in which the change should be promoted is a direction toward the center of the figure. This makes it possible to display not only the direction in which the change should be promoted, but also the degree of change that should be promoted.

(4) Furthermore, the process management support device of the present invention is characterized in that the predetermined standard is determined based on the accumulated data set and is the quality of the product with respect to the predetermined parameter. This makes it possible to display appropriate standards obtained from past process data for a specific management item.

(5) The process management support device of the present invention is also characterized in that the display control unit displays the product quality on the xy plane as the predetermined standard. This makes it possible to intuitively communicate to the operator what operating conditions are necessary to maintain the product quality.

(6) Furthermore, the process management support device of the present invention is characterized in that the product is clinker, which is an intermediate product of cement, and the predetermined parameters are the firing index and kiln current in the firing process. This makes it possible to display standards that are particularly important for maintaining product quality.

(7) The process management support device of the present invention is also characterized in that the display control unit displays, as the predetermined parameters, predetermined parameters of at least one of the process data from one cycle or more ago and the most recent process data. This makes it possible to clearly display the direction in which the process data is heading by arranging the past process data, the current process data, and the predicted value side by side. Note that while it is efficient to display only the most recent process data and the process data from one cycle or more ago, multiple process data from one cycle or more ago may be displayed.

(8) The process management support method of the present invention is a method for supporting process management in a product manufacturing plant, and is characterized by including the steps of acquiring process data representing the environment or settings in the target plant in real time, identifying predicted values for predetermined parameters in the acquired process data using a layer structure that has been deep-learned using an accumulated data set of the process data, and indicating the direction in which changes in the predetermined parameters should be encouraged by displaying the predetermined parameters and the predicted values in the acquired process data superimposed on a display of a predetermined criterion related to product quality. This makes it possible to communicate to the operator not only the current state of the manufacturing process but also future predicted states regarding product quality in an easy-to-understand manner.

(9) The process management support program of the present invention is a process management support program for a product manufacturing plant, and is characterized in that it causes a computer to execute the following processes: acquiring process data representing the environment or settings of a target plant in real time; identifying predicted values for predetermined parameters of the acquired process data using a layer structure that has been deep-learned using an accumulated data set of the process data; and indicating the direction in which changes to the predetermined parameters should be encouraged by displaying the predetermined parameters and the predicted values of the acquired process data superimposed on a display of a predetermined criterion related to product quality. This makes it possible to communicate to the operator not only the current state of the manufacturing process, but also future predicted states regarding product quality in an easy-to-understand manner.

(10) The system of the present invention is characterized by comprising a plant having equipment for heating continuously supplied raw materials, and the support device described in any one of (1) to (7) above. As a result, even future predicted conditions can be communicated to the operator in an easily understandable manner in a plant where raw materials are continuously supplied.

The present invention makes it possible to communicate to operators not only the current status of the manufacturing process but also future predicted status regarding product quality in an easy-to-understand manner.

1 is a schematic diagram showing a configuration of a process management system according to the present invention; 1 is a block diagram showing a configuration of a process management support device according to the present invention; 13 is a flowchart showing a process management support process; FIG. 1 is a diagram showing the layer structure of deep learning. FIG. FIG. 13 is a diagram showing the transition of predicted and measured values of kiln current. This is a diagram in which past, present and predicted values are plotted on an xy plane with two axes of kiln current and firing index.

The following describes an embodiment of the present invention.

[Configuration of the process management system]
Fig. 1 is a schematic diagram showing a process control system 10. The process control system 10 includes a plant 100 and a process control support device 200. In the example shown in Fig. 1, the plant 100 is a facility that performs a cement calcination process in a cement factory. The cement manufacturing process is divided into three processes: a raw material process, a calcination process, and a finishing process. Among them, the calcination process requires skilled operation because it simultaneously handles "powder,""viscousfluid," and "solid," has a long response time, and is subject to many disturbances.

Therefore, there is great significance in utilizing accumulated data and deep learning to improve operations. The present invention is suitable for the cement burning process, but is not limited to this, and can also be applied to steel, chemicals, and other manufacturing processes that involve operations while viewing multiple process data, including cement raw material processes and finishing processes.

The process management support device 200 is a device such as a PC that has at least a CPU and memory, and notifies the operator OP of an appropriate operation based on information from the plant 100. The process management support device 200 also accepts operations from the operator OP and instructs the plant 100 to carry out the operation.

[Plant configuration]
The plant 100 has a raw material feed line 110, a new suspension preheater 120, a rotary kiln 140, a kiln pre-burner 150, a cooler 170, and a duct 180 for extracting air to the calciner. The configuration of the plant 100 will be described below along the cement manufacturing process.

The raw materials for clinker, an intermediate product of cement production, mainly include limestone, clay, silica stone, and iron raw materials. In addition, recycled resources such as waste plastic, waste tires, waste wood, and sewage sludge are also used as raw materials and heat sources.

In the raw material process, raw materials are mixed for each cement type and dried and crushed in a crusher. These raw materials are continuously fed to the firing process and fired. At this time, the raw materials are fed through the raw material feed line 110, passed through the new suspension preheater 120, and are fired as they pass through the rotary kiln 140, becoming clinker, which then flows into the cooler 170.

In the cooler 170, the burnt clinker is rapidly cooled and its heat is recovered and used as combustion air, realizing a heat circulation process that reduces heat loss. The gas is air taken in by the cooler 170, passes through the rotary kiln 140 and each cyclone 121, and is discharged from the induction fan 123. The feed speed of the raw materials is controlled by the position of the raw material feed passage 110.

The new suspension preheater 120 includes cyclones 121 and a calciner 125, and each cyclone 121 separates the gas from the raw materials. The gas that enters the lowest cyclone 121 from the rotary kiln 140 exchanges heat with the raw materials as it passes through each cyclone 121, and is discharged outside the system by an induced draft fan 123. The discharged gas is also used to dry the raw materials and generate waste heat. The rotation speed of the induced draft fan 123 is controlled. The gas temperature of the four cyclones 121 at the top is monitored by a temperature sensor. The calciner 125 is equipped with a burner, and the temperature inside the furnace is controlled.

The raw materials fed into the rotary kiln 140 through the news suspension preheater 120 enter from the kiln end 143 and are heated to about 1450°C as they proceed to the kiln front 145. The _O2 and _NOx concentrations at the kiln end 143 are monitored. The kiln front burner 150 is installed at the kiln front 145 of the rotary kiln 140. During this burning process, the raw materials undergo gradual chemical changes, becoming clinker, a collection of hydraulic compounds and an intermediate product of cement, which is discharged from the kiln front 145. The current of the rotary kiln 140 is monitored, and the rotation speed of the rotary kiln 140 is controlled.

The clinker that has become a compound in the kiln enters the cooler through the drop port 148 and is rapidly cooled. The temperature at the drop port 148 is monitored. Between the cooler 170 and the calciner 125, a duct 180 is provided to extract air from the cooler to the calciner, and an air extraction damper 185 is provided in the duct 180 to adjust the amount of air extracted from the cooler. The opening of the air extraction damper 185 is controlled. The free lime in the clinker discharged from the cooler 170 is monitored by component analysis. In the finishing process, gypsum and admixtures are added to the clinker and it is crushed in a crusher to become the final product, cement.

The cement burning process mentioned above is not only difficult in itself, but it is also an interference process in which one operation affects multiple processes. Furthermore, fluctuations in the calorie content of recycled resources are a major disturbance, making operations even more difficult in recent years.

In the cement burning process, the operator performs operations based on information related to the process in accordance with standards established from the perspectives of the environment, quality, and energy conservation. The kiln current is one of the items monitored by the operator, and stabilizing it leads to stabilization of the production volume and quality of clinker. To achieve appropriate operation, it is important to understand the characteristics of the process. In addition, experienced operators understand the characteristics of the process and can perform pre-emptive operations that include predictions in their situation judgments. Up until now, efforts have been made to model processes using physical and statistical methods, but with the increase in recycled resources, it is difficult to maintain universality even in modeling. For these reasons, there are few feedforward elements based on predictions in current kiln operations, and feedback-centered operations are unavoidable. For these reasons, operation support using deep learning as in the present invention is suitable for cement manufacturing processes, especially the burning process. Note that the above example is one suitable example, and the monitored and controlled objects may be selected appropriately depending on the plant.

[Configuration of the Process Management Support Device]
Fig. 2 is a block diagram showing the configuration of a process management support device 200. As shown in Fig. 2, the process management support device 200 includes an information acquisition unit 210, a storage unit 220, a learning processing unit 240, a predicted value specification unit 250, a display control unit 255, a display unit 260, and an operation unit 270. The solid lines shown in Fig. 2 represent the transmission and reception of information for operation prediction, and the dashed dotted lines represent the transmission and reception of information for learning processing.

The information acquisition unit 210 acquires process data in the target plant in real time. The acquired process data is stored in the memory unit 220 as teacher data, and is sent to the prediction value identification unit 250. Some process data represents the environment or settings. As described above, the memory unit 220 accumulates process data necessary for learning by the prediction value identification unit 250. If data identical to the stored process data can be prepared by other means, accumulation in the memory unit 220 can be omitted.

Process data is a collection of multiple types of data that represent the environment and settings, and when these are interdependent, it is preferable to apply a layered structure that enables deep learning. In particular, in the cement manufacturing process, not only sensed information but also input from the operating terminal can affect multiple future process data. This makes it possible to transfer skills in judging changes in process data and making predictive judgments, which previously had to be relied on the individual experience and intuition of operators.

The predicted value is preferably data representing the future environment, and the number of minutes into the future the value may be determined based on the training data used.

The learning processing unit 240 receives process data and information on actual operations from the memory unit 220, and receives predicted values from the predicted value identification unit 250. It then learns to minimize an evaluation criterion (described below) from the actual values and predicted values, and updates the information that constitutes the hidden layer of the layered structure. In this way, the predicted value identification unit 250 learns from past process data in the plant 100 as teacher data.

The predicted value identification unit 250 identifies a predicted value for the acquired process data using a deep learning layer structure. This enables an operator without experience, intuition, or other skills to manage the process with the same foresight judgment as an experienced operator.

The display control unit 255 indicates the direction in which changes to the process data should be encouraged by displaying specific parameters from the acquired process data and their predicted values overlaid on the display of specific criteria related to product quality. This makes it possible to clearly communicate to the operator not only the current state of the manufacturing process with respect to product quality, but also the predicted future state. The operator can then utilize the provided information in their operations to improve the manufacturing process.

In this case, it is preferable to display a predetermined standard, the acquired process data, and the predicted value on an xy plane with xy axes formed by two specific types of process data. This makes it possible to intuitively and easily display the direction in which the process data is heading.

As a predetermined standard, for example, it is preferable to display the product quality three-dimensionally on an xy plane using another parameter. This makes it possible to intuitively communicate to the operator what operating conditions are necessary to maintain the product quality. An example of such a three-dimensional display will be described later.

The specified standard is a figure enclosed by line segments on a specified xy plane, and the direction in which the change should be encouraged is preferably a direction toward the center of the figure. This makes it possible to display not only the direction in which the change should be encouraged, but also the degree of change that should be encouraged.

The predetermined criteria are determined based on accumulated data sets and are preferably product quality for two specific types of data. This allows for the display of reasonable criteria obtained from past process data for a specific control item.

In addition, it is preferable to display the data from one cycle ago and the most recent data as a specified parameter of the acquired process data. This allows past data, current data, and predicted values to be displayed side by side, making it possible to clearly display the direction in which future data is heading.

The display unit 260 includes an output device such as a display and an output circuit, and displays the identified predicted value. The operation unit 270 includes an input device such as a keyboard and a mouse and an input circuit, and receives operations by the operator OP and instructs the plant 100 to perform the operations.

[Operation of the Process Management Support Device]
The operation of the process management support device 200 configured as above will be described. Fig. 3 is a flowchart showing the process management support process. As shown in Fig. 3, first, process data is acquired in real time (step S1), and a predicted value is identified using a layer structure of deep learning (step S2).

The predicted values include data and measured values indicating future states. The measured values include, for example, kiln current in a cement manufacturing process. Then, the current values and predicted values are displayed on an xy plane with two specific axes for a predetermined standard related to the quality of the product (step S3). The current values and predicted values are expressed by predetermined parameters of the process data, such as the kiln current and the firing index. The above display conveys to the operator OP the future trends in the process data and the direction in which the process data should be changed. The operator OP refers to the transmitted information and inputs operations to the process management support device 200, and the process management support device 200 instructs the plant 100 on the accepted operations (step S4). In this way, process management is supported, and even non-experts can easily operate the system.

Next, the operation of the learning process will be described. Learning is performed by updating the deep learning layer structure from the actual values and predicted values. Specifically, the weights within the deep learning layer structure are adjusted. Basically, iterative learning is performed on data for a period such as one or two years, but sequential learning may also be performed. Note that in order to learn the relationship between the current measurement value and the measurement value to be predicted, it is preferable to shift the entire data of the actual measurement value of the measurement value to be predicted by a predetermined amount of time toward the past.

In the above example, the target is a cement factory that produces cement as a product, but the above device is also suitable in other cases, for example, when the plant includes equipment that heats a continuously supplied supply of raw materials, the process data includes the temperature of the equipment, and the set value of the control terminal includes the feed rate of raw materials or fuel into the equipment.

In other words, the temperature of a continuously operating facility (such as a furnace) is determined by many factors, and the rate at which raw materials or fuel are fed into the facility has a variety of effects on the conditions within the facility. In process management where such complex interactions between elements occur, support devices that use neural networks like those described above are effective.

[Layer structure of deep learning]
Next, the layer structure of deep learning will be described. Fig. 4 is a diagram showing the layer structure of deep learning.

Deep learning has multiple hidden layers between the input layer and output layer. It has a similar layer structure to a neural network, but the procedure for acquiring the features of each node in the hidden layer differs from that of a neural network. In a neural network, hidden layers are trained using backpropagation, but deep learning trains so that the input and output of the hidden layers match (autoencoder).

For example, the autoencoder can be trained to minimize the evaluation criterion shown in formula (1). In the evaluation criterion, in addition to minimizing the Mean Square Error (MSE), a term is provided that constrains the weights of the autoencoder and the sparsity. Here, x and x with a hat symbol represent the actual measured value and predicted value of the prediction target, the variable N represents the number of observation examples, and D represents the number of variables.

The Ω_weights term in formula (1) is an L2 regularization term shown in formula (2), and overlearning can be suppressed by introducing it into the evaluation criterion, where L is the number of hidden layers and w is the weight.

In addition, the Ω_sparsity term in formula (1) is a sparse regularization term shown in formula (3). By introducing this term into the evaluation criterion, the sparsity of the hidden layer can be increased. ρ and ρ with a hat symbol indicate the output target value and the activated output value of each element.

[Example]
In fact, in a cement factory, the accumulated data was utilized, and a system using a deep learning layer structure was used to output the predicted values of the kiln current and the firing index in real time. The process data used includes input data (measured values) such as the calciner temperature, the supply rate of the pulverized coal before the kiln, the kiln rotation speed, the feed rate of the raw material, the induction fan rotation speed, the extraction damper opening, the kiln current, the gas temperature and the raw material temperature of each cyclone, the drop-out clinker temperature, the kiln end _O2 concentration, the kiln end NOx concentration, and the free lime in the clinker, as well as input data (set values) such as the calciner temperature control, the supply rate of the pulverized coal before the kiln, the kiln rotation speed, the feed rate of the raw material, the induction fan rotation speed, and the extraction damper opening. In the embodiment, past process data was used during learning. In addition, the input variables in the implemented deep learning layer structure are 183, the number of hidden layers is 3, and the output variable is 1.

Two years' worth of data from 2017 to 2018 were prepared as learning data, and learning was performed in three patterns: 2017 data only, 2018 data only, and 2017-18 data. These were then applied to a period of data in 2019 and verification was performed. The preprocessing, evaluation method, and verification results are explained below. Note that while it is desirable for learning data to include a variety of operating conditions, spike signals and other values that are thought to affect learning were deemed unnecessary and were removed from each process value.

The LAG at which the cross-correlation coefficient R for the predicted value and the actual measured value samples is maximized was determined as the prediction lead, and the coefficient of determination ^R2 at that time was calculated. LAG refers to the amount of shift in the actual measured value data. Here, a negative LAG means a lead. The sampling period was 1 point/min, the number of sample points was 4320 points (1440 points/day x 3 days), and the number of sample periods was 10 periods.

The above evaluation method was used for verification. Figure 5 shows the driving results. As a result of learning using one year's worth of data from 2017 and 2018, there were cases where predictions were possible and cases where they were not possible depending on the sample period, resulting in a lack of stability. In contrast, the result of learning using two years' worth of data from 2017 to 2018 showed a performance that was 16 minutes ahead on average in all sample periods and was able to handle a variety of driving situations. This shows that in order to make full use of deep learning, the quantity of data as well as the quality of data are important.

The two years of learning from 2017-18, when good results were obtained, were then applied to the actual machine to predict kiln current in real time. Features that output future kiln current and firing index were extracted from the process data, and future kiln current was output (predicted) based on the input data. Figure 6 shows the trends in predicted and actual measured values of kiln current. As shown in Figure 6, it can be seen that the predicted value is ahead of the actual measured value of kiln current. The same accuracy as the verification results was confirmed in the actual machine.

To encourage operators to take proactive action, it is possible not only to display the predicted value of kiln current as a trend, but also to learn and predict other firing indicators in the same way as the kiln current, and to replace them with a form that links the predicted values of kiln current and other firing indicators with quality indicators. Figure 7 is a diagram in which past, current, and predicted values are plotted on a two-axis xy plane of kiln current and firing indicators. The diagram visually displays the relationship (contour lines) between each indicator, and also plots past and current actual values and predicted values on a two-axis xy plane. This makes it possible to grasp changes in operating conditions and target ranges at a glance.

As shown in Figure 7, the current kiln current and firing index and their predicted values are represented by double circles and stars, allowing you to see future trends. In addition, the kiln current and firing index from the past (one cycle ago, for example, five minutes ago) can also be displayed to make the trend even easier to understand.

In the example shown in Figure 7, the quality of the xy-plane product is displayed three-dimensionally in shades on the xy-plane with the xy axes formed by two specific types of process data (kiln current and firing index). These quality indications C1 to C6 are determined based on accumulated data sets and are appropriate standards obtained from past process data for specific management items. Quality indications C1 to C2 indicate a state where the product is sufficiently fired, but is fired too much and heat is being wasted. Quality indications C5 to C6 indicate a state where the firing is insufficient. It can be shown that the kiln current and firing index should be maintained in the range of quality indications C3 to C4.

In addition to shading, quality can also be displayed using different colors, or quality can be displayed on the z-axis, with the xyz space viewed from an angle. Displaying it in this way makes it possible to intuitively communicate to the operator how to control the kiln current and firing indicators to maintain product quality. Firing indicators include measured values that change depending on the firing conditions, such as the temperature, pressure, and gas analysis values during the firing process, as well as calculated values such as the amount of heat held by the preheater and kiln.

The predetermined criterion may be a closed figure formed by line segments on a specified xy plane. In the example shown in Figure 7, a square frame is shown. Product quality can be maintained by changing the kiln current and firing index so that the product moves toward the center of the square.

10 Process management system 100 Plant 110 Raw material inlet 120 New suspension preheater 121 Cyclone 123 Induction fan 125 Calciner 140 Rotary kiln 143 Kiln butt 145 Kiln front 148 Drop opening 150 Kiln front burner 170 Cooler 180 Duct 185 Extraction damper 200 Support device 210 Information acquisition unit 220 Memory unit 240 Learning processing unit 250 Prediction value identification unit 255 Display control unit 260 Display unit 270 Operation unit OP Operators C1 to C6 Quality display

Claims (8)

A support device for process management in a product manufacturing plant, comprising:
an information acquisition unit that acquires process data representing an environment or setting in a target plant in real time;
a prediction value specifying unit that specifies a prediction value for a predetermined parameter of the acquired process data by using a layer structure that is deep-learned using an accumulated data set of the process data;
a display control unit that displays a predetermined parameter and the predicted value of the acquired process data superimposed on a display of a predetermined standard related to product quality, thereby indicating a direction in which the predetermined parameter should be changed ;
displaying the predetermined criterion, a predetermined parameter of the acquired process data, and the predicted value on an xy plane defined by xy axes formed by the two types of the predetermined parameters;
An assistance device characterized in that, as the specified standard, the xy range is adjusted so that the center of a closed figure on the xy plane is the center of the area in which product quality should be maintained, thereby setting the direction in which the change should be promoted to be a direction toward the center of the figure . 2. The apparatus of claim 1 , wherein the predetermined criteria is determined based on the accumulated data set and is a quality of the product relative to the predetermined parameters. 3. The support device according to claim 2 , wherein the display control unit displays, as the predetermined criterion, product quality on the xy plane. The product is clinker, which is an intermediate product of cement,
4. The support device according to claim 1 , wherein the predetermined parameters are a firing index and a kiln current in the firing process. 5. The support device according to claim 1, wherein the display control unit displays, as the predetermined parameters, predetermined parameters of at least one of the process data from one period earlier and the most recent process data. A method for supporting process management in a product manufacturing plant, comprising:
acquiring process data representative of an environment or setting in a target plant in real time;
identifying predicted values for predetermined parameters of the acquired process data using a layer structure that is deep-learned using the accumulated data set of the process data;
and displaying a predetermined parameter and the predicted value of the acquired process data superimposed on a display of a predetermined criterion related to product quality, thereby indicating a direction in which a change in the predetermined parameter should be promoted ;
displaying the predetermined criterion, a predetermined parameter of the acquired process data, and the predicted value on an xy plane defined by xy axes formed by the two types of the predetermined parameters;
The support method is characterized in that the direction in which the change should be promoted is a direction toward the center of the figure by adjusting the xy range so that the center of the closed figure on the xy plane is the center of the area in which product quality should be maintained as the specified standard . A support program for process management in a product manufacturing plant, comprising:
acquiring process data representative of an environment or setting in a target plant in real time;
identifying a predicted value for a predetermined parameter of the acquired process data using a layer structure that is deep-learned using an accumulated data set of the process data;
a process of displaying a predetermined parameter and the predicted value of the acquired process data superimposed on a display of a predetermined standard related to product quality, thereby indicating a direction in which the predetermined parameter should be changed ;
displaying the predetermined criterion, a predetermined parameter of the acquired process data, and the predicted value on an xy plane defined by xy axes formed by the two types of the predetermined parameters;
An assistance program characterized by adjusting the xy range so that the center of a closed figure on the xy plane is the center of an area in which product quality should be maintained as the specified standard, thereby setting the direction in which the change should be promoted to a direction toward the center of the figure . A plant having a facility for heating a continuously supplied raw material;
A system comprising: a support device according to any one of claims 1 to 5 .

Images