Search

A frame model is described, which underlies moisture content operational assessment system functioning during its drying in drum drying installations. Drying plant equipment includes a variety of sensors that monitor the drying process. Failure of one or several sensors can have a negative impact on the drying process information content, but won't bring any effect on the dried alcohol stillage quality. Therefore, sensors groups are identified in a way to correspond states classes: failure-free operation, soft failures, medium failures, and sensors' hard failures. The frames of the created model describe equipment drying plant functioning states classes, and specific properties of each class are reflected in the slots containing the attached procedures for neural networks implementation that allow moisture content material estimation. The frame model allows to create adaptive sensors for material's moisture content operational assessment, taking into account equipment operation various states, making up the drying unit and its operating modes. The use of an adaptive sensor makes it possible to estimate alcohol stillage moisture content for all operating states with an error less than 2 %.

The article considers creation of a system for energy-saving dynamic modes control, invariant to various heat-technological devices. Heat-technological devices and their application classes are given. The features of managing such classes of objects are highlighted. The problem of energy-saving control with fixed intervals for time and control action is formulated. Two classes heat-technological devices energy-saving control problem solving examples are given: a roller-belt drying unit and an annealing furnace for magnetic circuits. For a roller-belt drying plant, the functional fuel consumption is minimized, and for the annealing magnetic circuits furnace, the functional is energy consumption. Minimizing energy costs problem takes into account heat-technological device functioning various states. The proposed energy-saving control system industry introduction allows to save from 5% up to 15% of energy in heat-technological devices dynamic modes operation.

The article describes a control information system development aimed to save energy in electric furnace magnetic cores annealing dynamic modes control. System’s control objects main features include high energy intensity, operation frequency and their insufficient isolation from the environment. Due to their design features, electric magnetic core annealing furnaces are continuous. As a result, significant heat losses to the environment occur. Regarding control actions analysis and synthesis stages, synthesizing variables are introduced, bringing control problem dimension reduction. In accordance with operating plurality states, energy-saving control functions are defined that allow energy-saving control action real-time synthesis. The developed information system for annealing magnetic cores electric furnaces heating process control allows to save from 5% up to 15% electricity, depending on the operation’s state.

Developing a control system for heat-technological devices according to manufactured products quality losses minimizing criteria, great attention should be paid to the experts’ accumulated experience. The article provides possible situations classification procedural model application example for pasty material drying process control in a five-section roller-belt drying unit. The absolute value of the moisture content of the material at the drying plant outlet is less than 1%. The proposed possible situations classification procedural model is based on the Dempster-Shafer method, which allows, on the numerical estimates basis experts’ confidence-building measures, to determine at the beginning of the process whether the required quality indicator could be achieved for the desired control parameters. Proposed drying process control procedural model usage made it possible to increase the probability of a high-quality product output up to ninety-eight percent.

During drying process control according to the manufactured products and productivity quality loss minimization criteria, an important task is rapid pasty materials moisture assessment. Due to a control object features number, the drying process, or the control process itself, the known measuring devices cannot be used to achieve the criteria formulated above. Therefore, a method is proposed for the paste material moisture rapid assessment during its drying. The method features neural networks, built on pasty materials drying process most influent parameters’ significant sample basis. The article provides the described method usage example implementing device. The developed method makes it possible to evaluate pasty material moisture during its drying in real-time with a relative error less than 2%, which makes rapid control action synthesis possible, aimed to minimize the desired criteria.

There are several control objects main features, among them are: high requirements for maintaining technological regimes and control object’s main operating parameters changes in its’ actual operation process. The control vector may contain components of two types – general-action components for the entire drying plant and private (local) components. The control device generates actions that take into account possible changes in functioning states due to the time lag presence between the control actions and the output variables. At the drying processes control actions analysis stage, plenty of functioning states for various tasks classes are determined. Neural networks were used to assess the moisture material content during its drying, production rules and fuzzy logic in the control action synthesis. The developed drying process control information system usage for a paste-like dye in roller-ribbon drying plants can increase high-quality products probability design up to 0.98 and increase overall productivity by 5%.