Padang
Idwar
1), Hayatul Muttaqin
2) 1) UNIDHA , Padang, IndonesiaE-mail: [email protected]
2)UNIDHA, Padang, Indonesia
E-mail: [email protected]
Abstract. Along with advances in information technology and artificial intelligence that continues to grow rapidly, have made
it possible to move from human expert system to a computer system, making it possible to solve problems that require expertise without having to bring in the experts. The expertise of an expert is transferred to a computer, knowledge that is stored in the computer, and users can consult on the computer for an advice, then the computer can take inference (infer, deduce, etc.) Like an expert, as knowledge of an expert has transferred into the computer. The design of an expert system used to diagnose disorders of the Diesel Power Plant (PLTD) using backward chaining. The purpose of this development could eventually help in diagnosing disorders of the diesel and the results are precise and accurate diagnosis, and their advice is expected to facilitate the handling of the operator and the management in taking action / decision. Application design support are used is a relational database in Microsoft Office Access 2007.
Keywords: Design System, Expert System, inference, Backward Chaining, diagnosing disorders
I. INTRODUCTION
Developments in information technology today runs very fast and plays an important role in a variety of ways. Computer is one important part in improving information technology. The ability of the computer to remember and store information can be used without having to depend on barriers like those of humans, such as the condition of hunger, thirst or emotional. By storing the information and a set of rules that adequate reasoning allows computers provide conclusions or take decisions that same quality invitation ability of a
particular scientific field experts.
Expert systems are computer-based systems that use knowledge, facts and reasoning techniques in solving the problem, which usually can only be resolved by an expert in a particular field. Martin and Oxman, [1]
The expert system is a field which is characterized by knowledge-based systems (Knowledge Base System), allowing the computer can think and make decisions on a set of rules. Ignizio, [2]
Expert systems are computer programs that are designed to replicate the problem-solving skills of an expert. Specialists are those who have the ability or understanding in the face of a problem. Through experience, an expert develop the skills that allowed him to solve the problems with good results and efficient. John Durkin, [3]
An expert system is a branch of artificial intelligence that uses knowledge-specific knowledge possessed by an expert to solve a particular problem. Giarratano and Riley, [4] In general, the expert system is a system that is trying to adopt human knowledge into a computer that is designed to model the ability to solve problems like an expert. Sri Kusumadewi, [5].
With this expert system, any layperson can solve the problem or just looking for a real quality information can only be obtained with the help of experts in the field. This expert system will also be able to help the activities of experts as assistants who are experienced and have the required knowledge. In preparation, the expert system combines the rules of inference (inference rules) with a certain knowledge base provided by one or more experts in a particular field. The combination of both of these are stored in the computer, which is then used in the decision-making process for the settlement of certain issues.
PT Semen Padang is a company founded on March 18, 1910 under the name NV Nederlandsch Indische Portland Cement Maatschappij (NV NIPCM) which is the first cement plant in Indonesia. Over time, PT Semen Padang undergoing a process of awakening through the rehabilitation and development of plant capacity Indarung I. Furthermore, the plant transformation of plant capacity expansion of wet process technology becomes dry process with the construction of the factory Indarung II, III, IV and V. The development of the plant capacity would it requires a source of power intake is greater. Therefore, PT Semen Padang build several power plants Alone (PS) as the supporting main electrical intake of the State Electricity Company (PLN). One Power Alone PT Semen Padang is Diesel Power Plant (PLTD). PLTD is a vital asset as a backup unit producing power if at any time - the main source of power intake of PLN disconnected. Observing the things above, the writer got the idea to take advantage of computer technology to design an expert system application that is used to assist in diagnosing disorders of the diesel. Designing an expert system application program is useful for diagnosing disorders of the diesel. Program expert
system application is received input / update of the technicians and several other sources to be used as the foundation for an expert to diagnose the disorder in diesel and diagnosis given by the application program expert system is expected to help engineers and management in taking action / decision which of course for the sake of a smooth production process. Based on the descriptions above, the authors are interested in doing research with the title Design Expert Systems Applications Using Backward Chaining method Diagnosing Disorders of the Power Disesel (PLTD) in PT Semen Padang. Problem Formulation is How to design an expert system for diagnosing disorders of the Diesel Power Plant (PLTD) PT Semen Padang along with suggestions to handle the expected allow the operator and the management in taking action / decision ?
This research aims to devise an expert system that can be further developed to provide quick information about the obstruction on the Diesel Power Plant (diesel) PT Semen Padang, along with suggestions to handle the expected allow the operator and the management in taking action/ decision.
II. METHOD
Backward chaining rules are processed when your program asks Pyke a question (i.e., asks Pyke to prove a specific goal). See Figure 1. Pyke will only use activated rule bases to do the proof.
Overview of "Backward-Chaining"
To do backward-chaining, Pyke finds rules whose then part matches the goal (i.e., the question). Once it finds such a rule, it tries to (recursively) prove all of the subgoals in the if part of that rule. Some of these subgoals are matched against facts, and others are subgoals for other backward-chaining rules. If all of the subgoals can be proven, the rule succeeds and the original goal is proven. Otherwise, the rule fails, and Pyke tries to find another rule whose then part matches the goal, and so on.
So Pyke ends up linking (or chaining) the if part of the first rule to the then part of the next rule.
Reviewing:
1. Pyke starts by finding a rule whose then part matches the goal.
2. Pyke then proceeds to process the if part of that rule. 3. Which may link (or chain) to the then part of another
rule.
Since Pyke processes these rules from then to if to then to if in the reverse manner that we normally think of using rules, it's called backward chaining.
To make this more clear, Pyke has you write your backward- chaining rules upside down by writing the then part first (since that's how it is processed).
"Use", "When" Rather than "Then", "If"
But then-if rules sound confusing, so Pyke uses the words
use and when rather than then and if. You can then read the
rule as "use" this statement "when" these other statements can be proven.
Fig.1. i.e., asks Pyke to prove a specific goal http://pyke.sourceforge.net
Backward Chaining method is a method of decision- making were also commonly used in expert systems. Backward chaining method is the opposite of Forward Chaining. Experiment facts or statements starting from the right (first THEN). In other words, reasoning starts from the first hypothesis and to test the truth of this hypothesis to look facts in the knowledge base. The search process by backward chaining method departs from right to left, from the provisional conclusion leads to the premise, this method is often called a goal-driven search is controlled by the given destination.
Backward Chaining method is a search strategy of Forward Chaining the opposite direction. The search process starts from the goal, that conclusion into the solution of problems faced. Inference engine looking to the rules in the knowledge base that conclusion is a solution to be achieved, then of the rules obtained, respectively conclusion Backward Chaining pathway that leads to that conclusion. If the information or the values of attributes that lead to that conclusion in accordance with the data given the conclusions of such a solution is sought, if it does not match then the conclusion is not the solution sought. Backward Chaining start the search process for a purpose so this strategy also called goal-driven. See figure 2.
Fig. 2. Backward chaining Proces http://www.cse.unsw.edu.au III. RESULTS AND DISCUSSION
Problem analysis
Expert systems are structured system with dynamic knowledge base. Existing knowledge on the expert system can increases, so should be added or removed without changing
the contents of the program as a whole. So the change is only done on the basis of knowledge alone.
Menggunanakan knowledge base rule-based reasoning, knowledge represented by using rules of the form: IF-THEN. This form is used when we have a number of expert knowledge on a particular issue, and the expert can solve the problem sequentially. In addition, this form is also used when needed explanation of the trace (step - step) the achievement of a solution.
Stages analysis of a system done before the design stage is done. Objective analysis of the implementation of a system is to find out the reasons why such a system is necessary, formulate the needs of the system to reduce the excess resources as well as help plan the scheduling of the establishment of the system, so that the functions contained
within the system is working optimally.
One of the main elements that should be considered in the analysis phase of this system is a software problem, because the software that is used should be appropriate to the problem formulated.
Disorders Diagnosis PLTD
To design the knowledge base, the data obtained in consultation with the experts and books on diesel power plants. The data are designed such based on the relationships problems related to one another. Information obtained from sources it is known some PLTD common disorders that often occur. Here are the types of interference on PLTD PT Semen Padang:
TABLE I
DISORDERS MACHINE CAN NOT START
No Indicator Facts Disorders Handling Solutions
a. Generator in a state of lock / unlocked
Shaft does not
rotate Help with turning gear to rotate the shaft b. Pressure
water tank
Pressure water tank is empty
Turn on the compressor to fill air c. accessories
generator
Indicator lights of motor / pump inactive
Check the electrical pumps and mechanical contact d. Governor
machine Governor can not increase speed Check Switch, contact the electrical parts for repair
e. Handle fuel Supply fuel to the engine no
Check motors, pipes and fuel filter
TABLE II
DISORDERS CAN NOT PARALLEL MACHINES No. Indicator Facts Disorders Handling
Solutions
a. engine speed Engine speed does not qualify parallel Checks RPM (Rotation per minute) machine or a panel with tachometer b. excitation There is no voltage Reset Switch excitation ON, if not functioning electrical contact c. frequency machine The frequency of the engine and the system has not been the same
Set the engine RPM to match the frequency of the engine and the system d. Synchronization Needle synchronization inactive Set the synchronous switch ON for easy parallel. e. CB parallel The machine
does not generate power
Check the main CB panel space, set in the ON position TABLE III
DISTURBANCE FULL SPEED NOT BE ACHIEVED No. Indicator Facts
Disorders Handling Solutions
a. Suitability of fuel
Fuel instead of diesel
Replace fuel with diesel b. The fuel in tank mixes There was water on the surface of the solar
Suck the fuel mix
c. A fuel filter Fuel supply is
not smooth Clean the pipe and the fuel filter d. oil pressure Engine oil
pressure to less Checks if the blockage, and check the oil pressure pump motor TABLE VI
DISORDERS DOWN SPEED AND DEAD MACHINES No. Indicator Facts Disorders Handling Solutions
a. Availability
of fuel Fuel runs out Coordination with the addition of a warehouse for solar b. The fuel in
tank mixes
There was water on the surface of the solar
Suck water blended fuel
c. A fuel filter Fuel supply is not
smooth Clean the pipe and the fuel filter d. Supply of
fuel to the engine
The engine speed
TABLE V
PARALLEL DISORDERS ARRIVED DISCONNECTED No
. Indicator Facts Disorders Handling Solutions
a. The resulting voltage
The voltage is too high or too low
Set the voltage switches in the normal position b. The
resulting stream
Current is too high or too low
Set the voltage switches in the normal position c. frequency
system The frequency of the system is unstable
Coordinate with the substation if any additions or decrease the load d. Indication/s ignal from the generator There was damage to the machine Checks damages pursuant indica- tion lights are lit on the panel
Context Diagram
Is a global design tools that show the system in general and the parts of subsystems - subsystems involved in the system as a whole, linkages and interactions between subsystems designed himself, See Fig.3
Fig. 3. Context Diagram
Data Flow Diagrams (DFD)
Data Flow Diagrams (DFD) is used to describe the flow of information and the process of moving data from data entry to output. Data Flow Diagram allows users who are less mastered the field of computers to understand the system to be worked on or developed.
Entity Relationship Diagram (ERD)
Entity Relationship Diagram is used to model the data structures and relationships between data. Entity relationship diagrams used a number of notations and symbols to describe the structure and relationships between data.
In the system described several entities and attributes that are required as well as the relationships each designed for this application. For more details can be seen in Figure 4 below:
Fig.4. Entity Relationship Diagram
System implementation
Implementation of the system is the final stage of the development process after the application system through the design stage. In order for the process of implementation of the
software can work perfectly, the first such software should be tested to identify the weaknesses and errors are then evaluated.
Backtracking Implementation System, see table 6,7,8,9,10
TABLE VI
TABLE SYSTEM DISORDERS MACHINES CAN NOT START Questions answer fact handling Solution
a. Is the engine shaft can rotate? Yes No • shaft can rotate
• shaft does not rotate
• Shaft OK • Help with turning gear to rotate the shaft b. Is Air Pressure tank empty? Yes No
• The water tank is empty pressure • The tank holds wind • Turn on the compressor to fill air • Air Pressure OK c. Are all acces sories function generator Yes No • Indicator Lights motors / pumps active • Indicator Lights motors / pumps inactive • Accessories OK • Check electrical pumps and mechanical contact d. Is Governor machine function? Yes No • Governor can increase speed • Governor can not increase speed • Governor OK • Check Switch, contact the electrical parts for repair e. Do handle fuel in the ON state? Yes No • suply of fuel to the engine smoothly • suply of fuel to the engine no • Handle fuel OK • Check the motor, pipes and fuel filter
TABLE VII
SYSTEM DISORDERS CAN NOT PARALLEL MACHINES
Questions answer fact handling Solution
a.Is the machine has a maximum speed? Yes No • The engine speed max • The engine speed does not qualify parallel
• The engine speed is OK • Check RPM (Rotation per minute) machine or a panel with tachometer. b.Do excitation in the ON state? Yes No • The engine speed max • The engine speed does not qualify parallel
• Excitation OK
• Reset Switch excitation ON, if not functioning electrical contact c.Is the frequency of the engine and the same system? Yes No • The frequency of 50.00 Hz • The frequency of the engine and the system has not been the same
• Frequency OK • Set the engine RPM to match the frequency of the engine and the system. d. Is already synchronou s generator? Yes No • Needle active sync •Needles synchronization inactive • Synchronization OK • Set the synchronous switch ON for easy parallel. e. Is the CB parallel in the ON state? Yes No • Top CB ON • The machine does not generate power • CB parallel OK • Check the main CB panel space, set in the ON position
TABLEL VIII
TABLE SYSTEM DISORDERS FULL SPEED NOT BE ACHIEVED Questions answer fact handling Solution
a.Is the fuel suitable Yes No • Diesel fuel • Fuel is not diesel
. • Fuel OK • Replace fuel with diesel b.Is ditanki fuel mixed with water? Yes
No • There is water on the surface of the solar
• Fuel clean
• Suck fuel mixed with
water. • Fuel OK c.Is the fuel filter good? Yes No • fuels to cleaner engines • suply fuel noncurrent • Filter OK • Clean the pipe and the fuel filter d.Is the oil pressure normal? Yes No • The indicator on the normal panel • engine oil pressure to less
• oil pressure OK • Check if the blockage, and check the oil pressure pump motor
TABLE IX
TABLE SYSTEM DISORDERS DOWN SPEED AND DEAD MACHINES Questions answer fact handling Solution
a.Is enough fuel in the tank.? Yes No • Fuel pretty
• Fuel runs out • Fuel OK • Coordinate with the warehouse to the addition of solar b.Is ditanki fuel mixed with water? Yes No • There is water on the surface of the solar • Fuel clean
• Suck fuel mixed with water. • Fuel OK c.Is the fuel
filter good? Yes No
• fuels to cleaner engines • suply fuel noncurrent
• Filter OK • Clean the pipe and the fuel filter d. Is the fuel supply to the engine normal? Yes No • The engine speed is stable • The engine speed fluctuation • Supply of fuel OK • Check the fuel pump
TABLE X
TABLE PARALLEL SYSTEM DISORDERS ARRIVE DISCONNECTED Questions answer fact handling Solution
a. Is the generated voltage stable? Yes No • Voltage Stable • Voltage too high or too low
• Voltage OK • Set the voltage switches in the normal position b. Is the current generated is stable? Yes No • The current stable • Flow is too high or too low
• Flow OK • Set the voltage switches in the normal position c. Is the frequency of the system stable? Yes No • Frequency stable • Frequency system unstable • Frequency OK • Coordinate with the substation if any additions or decrease the load. d. Is there any indication / signal from the generator? Yes No • There is damage to the machine • Generator normal • Checks damages pursuant indication lights are lit on the panel
• Generator OK
Implementation Interface / Interface
a.
Main Menu page (Figure 5)On the main menu there are two buttons namely the diagnosis and expert admins, to level users can only use the facilities