Apparatus and method for assisting grinding machine

An assistance apparatus for assisting at least one grinding machine including a controller, the assistance apparatus comprising:

• a status information acquiring section that acquires a grinding condition as a status information, the grinding condition including set states associated with a plurality of movement command data that are controllable with the controller of the at least one grinding machine;
• an evaluation result acquiring section that acquires evaluation results of a plurality of evaluation objects that are obtained under the grinding condition, the evaluation objects being predetermined;
• a reward calculating section that calculates a reward for the status information based on the evaluation results;
• a value function storing section that stores a value function which is produced in a reinforcement learning that is based on the status information and the reward;
• a policy storing section that stores a policy which is obtained from the value function, the policy being configured to adjust the movement command data which corresponds to the status information such that the evaluation results are optimum;
• an action determining section that determines the movement command data to be adjusted and an adjustment amount at which said movement command data is adjusted, from among candidates of the plurality of movement command data that are adjustable, based on the status information and the policy;
• an action information outputting section that is configured to output contents determined by the action determining section to the controller, the contents including an action information;
• a value function updating section that is configured to update the value function stored in the value function storing section, based on an adjusted status information that is obtained by adjusting the status information based on the action information, and the reward for the adjusted status information;
• a policy updating section that updates the policy based on the updated value function;
• an evaluation result estimation model storing section that stores an evaluation result estimation model of indicating a relationship between the status information and the evaluation result;
• and
• an evaluation result estimating section that estimates the evaluation result which is obtained under the adjusted grinding conditions, based on the status information and the evaluation result estimation model,
• wherein the evaluation objects include at least one of:
• a grinding quality of a workpiece on which a grinding process is performed by the at least one grinding machine;
• a surface condition or a wear amount of a grinding wheel after the grinding process;
• and
• a process time that is required to perform the grinding process of the workpiece,
• wherein the evaluation result acquiring section is configured to acquire the estimated evaluation result as the evaluation result, the estimated evaluation result being estimated by the evaluation result estimating section,
• wherein the evaluation result estimation model storing section stores a thermal displacement amount calculation model as the evaluation result estimation model, the thermal displacement amount calculation model indicating a relationship between the status information and the thermal displacement amount of a structural member of the at least one grinding machine, and
• wherein the evaluation result estimating section calculates the wear amount of the grinding wheel based on the thermal displacement amount and the status information, the thermal displacement amount being calculated based on the status information and the thermal displacement amount calculation model.